-
Exploring Scaling Trends in LLM Robustness
Authors:
Nikolaus Howe,
Michał Zajac,
Ian McKenzie,
Oskar Hollinsworth,
Tom Tseng,
Pierre-Luc Bacon,
Adam Gleave
Abstract:
Language model capabilities predictably improve from scaling a model's size and training data. Motivated by this, increasingly large language models have been trained, yielding an array of impressive capabilities. Yet these models are vulnerable to adversarial prompts, such as "jailbreaks" that hijack models to perform undesired behaviors, posing a significant risk of misuse. Prior work indicates…
▽ More
Language model capabilities predictably improve from scaling a model's size and training data. Motivated by this, increasingly large language models have been trained, yielding an array of impressive capabilities. Yet these models are vulnerable to adversarial prompts, such as "jailbreaks" that hijack models to perform undesired behaviors, posing a significant risk of misuse. Prior work indicates that computer vision models become more robust with model and data scaling, raising the question: does language model robustness also improve with scale? We study this question empirically, finding that larger models respond substantially better to adversarial training, but there is little to no benefit from model scale in the absence of explicit defenses.
△ Less
Submitted 26 July, 2024; v1 submitted 25 July, 2024;
originally announced July 2024.
-
Error mitigation with stabilized noise in superconducting quantum processors
Authors:
Youngseok Kim,
Luke C. G. Govia,
Andrew Dane,
Ewout van den Berg,
David M. Zajac,
Bradley Mitchell,
Yinyu Liu,
Karthik Balakrishnan,
George Keefe,
Adam Stabile,
Emily Pritchett,
Jiri Stehlik,
Abhinav Kandala
Abstract:
Pre-fault tolerant quantum computers have already demonstrated the ability to estimate observable values accurately, at a scale beyond brute-force classical computation. This has been enabled by error mitigation techniques that often rely on a representative model on the device noise. However, learning and maintaining these models is complicated by fluctuations in the noise over unpredictable time…
▽ More
Pre-fault tolerant quantum computers have already demonstrated the ability to estimate observable values accurately, at a scale beyond brute-force classical computation. This has been enabled by error mitigation techniques that often rely on a representative model on the device noise. However, learning and maintaining these models is complicated by fluctuations in the noise over unpredictable time scales, for instance, arising from resonant interactions between superconducting qubits and defect two-level systems (TLS). Such interactions affect the stability and uniformity of device performance as a whole, but also affect the noise model accuracy, leading to incorrect observable estimation. Here, we experimentally demonstrate that tuning of the qubit-TLS interactions helps reduce noise instabilities and consequently enables more reliable error-mitigation performance. These experiments provide a controlled platform for studying the performance of error mitigation in the presence of quasi-static noise. We anticipate that the capabilities introduced here will be crucial for the exploration of quantum applications on solid-state processors at non-trivial scales.
△ Less
Submitted 5 July, 2024; v1 submitted 2 July, 2024;
originally announced July 2024.
-
Optical Control of Adaptive Nanoscale Domain Networks
Authors:
Marc Zajac,
Tao Zhou,
Tiannan Yang,
Sujit Das,
Yue Cao,
Burak Guzelturk,
Vladimir Stoica,
Mathew Cherukara,
John W. Freeland,
Venkatraman Gopalan,
Ramamoorthy Ramesh,
Lane W. Martin,
Long-Qing Chen,
Martin Holt,
Stephan Hruszkewycz,
Haidan Wen
Abstract:
Adaptive networks can sense and adjust to dynamic environments to optimize their performance. Understanding their nanoscale responses to external stimuli is essential for applications in nanodevices and neuromorphic computing. However, it is challenging to image such responses on the nanoscale with crystallographic sensitivity. Here, the evolution of nanodomain networks in (PbTiO3)n/(SrTiO3)n supe…
▽ More
Adaptive networks can sense and adjust to dynamic environments to optimize their performance. Understanding their nanoscale responses to external stimuli is essential for applications in nanodevices and neuromorphic computing. However, it is challenging to image such responses on the nanoscale with crystallographic sensitivity. Here, the evolution of nanodomain networks in (PbTiO3)n/(SrTiO3)n superlattices was directly visualized in real space as the system adapts to ultrafast repetitive optical excitations that emulate controlled neural inputs. The adaptive response allows the system to explore a wealth of metastable states that were previously inaccessible. Their reconfiguration and competition were quantitatively measured by scanning x-ray nanodiffraction as a function of the number of applied pulses, in which crystallographic characteristics were quantitatively assessed by assorted diffraction patterns using unsupervised machine-learning methods. The corresponding domain boundaries and their connectivity were drastically altered by light, holding promise for light-programmable nanocircuits in analogy to neuroplasticity. Phase-field simulations elucidate that the reconfiguration of the domain networks is a result of the interplay between photocarriers and transient lattice temperature. The demonstrated optical control scheme and the uncovered nanoscopic insights open opportunities for remote control of adaptive nanoscale domain networks.
△ Less
Submitted 24 June, 2024;
originally announced June 2024.
-
Quantum embedding study of strain and charge induced Stark effects on the NV$^{-}$ center in diamond
Authors:
Gabriel I. López-Morales,
Joanna M. Zajac,
Johannes Flick,
Carlos A. Meriles,
Cyrus E. Dreyer
Abstract:
The NV$^{-}$ color center in diamond has been demonstrated as a powerful nanosensor for quantum metrology, due to the sensitivity of its optical and spin properties to external electric, magnetic, and strain fields. In view of these applications, we use quantum embedding to derive a many-body description of strain and charge induced Stark effects on the NV$^{-}$ center. We quantify how strain long…
▽ More
The NV$^{-}$ color center in diamond has been demonstrated as a powerful nanosensor for quantum metrology, due to the sensitivity of its optical and spin properties to external electric, magnetic, and strain fields. In view of these applications, we use quantum embedding to derive a many-body description of strain and charge induced Stark effects on the NV$^{-}$ center. We quantify how strain longitudinal to the axis of NV$^{-}$ shifts the excited states in energy, while strain with a component transverse to the NV axis splits the degeneracies of the $^{3}E$ and $^{1}E$ states. The largest effects are for the optically relevant $^{3}E$ manifold, which splits into $E_{x}$ and $E_{y}$ with transverse strain. From these responses we extract strain susceptibilities for the $E_{x/y}$ states within the quasi-linear regime. Additionally, we study the many-body dipole matrix elements of the NV$^{-}$ and find a permanent dipole 1.64 D at zero strain, which is $\sim 30 \%$ smaller than that obtained from recent density functional theory calculations. We also determine the transition dipole between the $E_{x}$ and $E_{y}$ and how it evolves with strain.
△ Less
Submitted 11 June, 2024;
originally announced June 2024.
-
Hybrid Data Management Architecture for Present Quantum Computing
Authors:
Markus Zajac,
Uta Störl
Abstract:
Quantum computers promise polynomial or exponential speed-up in solving certain problems compared to classical computers. However, in practical use, there are currently a number of fundamental technical challenges. One of them concerns the loading of data into quantum computers, since they cannot access common databases. In this vision paper, we develop a hybrid data management architecture in whi…
▽ More
Quantum computers promise polynomial or exponential speed-up in solving certain problems compared to classical computers. However, in practical use, there are currently a number of fundamental technical challenges. One of them concerns the loading of data into quantum computers, since they cannot access common databases. In this vision paper, we develop a hybrid data management architecture in which databases can serve as data sources for quantum algorithms. To test the architecture, we perform experiments in which we assign data points stored in a database to clusters. For cluster assignment, a quantum algorithm processes this data by determining the distances between data points and cluster centroids.
△ Less
Submitted 18 March, 2024; v1 submitted 12 March, 2024;
originally announced March 2024.
-
Solving Distributed Flexible Job Shop Scheduling Problems in the Wool Textile Industry with Quantum Annealing
Authors:
Lilia Toma,
Markus Zajac,
Uta Störl
Abstract:
Many modern manufacturing companies have evolved from a single production site to a multi-factory production environment that must handle both geographically dispersed production orders and their multi-site production steps. The availability of a range of machines in different locations capable of performing the same operation and shipping times between factories have transformed planning systems…
▽ More
Many modern manufacturing companies have evolved from a single production site to a multi-factory production environment that must handle both geographically dispersed production orders and their multi-site production steps. The availability of a range of machines in different locations capable of performing the same operation and shipping times between factories have transformed planning systems from the classic Job Shop Scheduling Problem (JSSP) to Distributed Flexible Job Shop Scheduling Problem (DFJSP). As a result, the complexity of production planning has increased significantly. In our work, we use Quantum Annealing (QA) to solve the DFJSP. In addition to the assignment of production orders to production sites, the assignment of production steps to production sites also takes place. This requirement is based on a real use case of a wool textile manufacturer. To investigate the applicability of this method to large problem instances, problems ranging from 50 variables up to 250 variables, the largest problem that could be embedded into a D-Wave quantum annealer Quantum Processing Unit (QPU), are formulated and solved. Special attention is dedicated to the determination of the Lagrange parameters of the Quadratic Unconstrained Binary Optimization (QUBO) model and the QPU configuration parameters, as these factors can significantly impact solution quality. The obtained solutions are compared to solutions obtained by Simulated Annealing (SA), both in terms of solution quality and calculation time. The results demonstrate that QA has the potential to solve large problem instances specific to the industry.
△ Less
Submitted 11 March, 2024;
originally announced March 2024.
-
Imprinting of Antiferromagnetic Vortex States in NiO-Fe Nanostructures
Authors:
M. Ślęzak,
T. Wagner,
V. K. Bharadwaj,
O. Gomonay,
A. Kozioł-Rachwał,
T. O. Menteş,
A. Locatelli,
M. Zając,
D. Wilgocka-Ślęzak,
P. Dróżdż,
T. Ślęzak
Abstract:
Magnetic vortices are topological spin structures frequently found in ferromagnets, yet novel to antiferromagnets. By combining experiment and theory, we demonstrate that in a nanostructured antiferromagnetic-ferromagnetic NiO(111)-Fe(110) bilayer, a magnetic vortex is naturally stabilized by magnetostatic interactions in the ferromagnet and is imprinted onto the adjacent antiferromagnet via inter…
▽ More
Magnetic vortices are topological spin structures frequently found in ferromagnets, yet novel to antiferromagnets. By combining experiment and theory, we demonstrate that in a nanostructured antiferromagnetic-ferromagnetic NiO(111)-Fe(110) bilayer, a magnetic vortex is naturally stabilized by magnetostatic interactions in the ferromagnet and is imprinted onto the adjacent antiferromagnet via interface exchange coupling. We use micromagnetic simulations to construct a corresponding phase diagram of the stability of the imprinted antiferromagnetic vortex state. Our in depth analysis reveals that the interplay between interface exchange coupling and the antiferromagnet magnetic anisotropy plays a crucial role in locally reorienting the Néel vector out-of-plane in the prototypical in-plane antiferromagnet NiO and thereby stabilizing the vortices in the antiferromagnet.
△ Less
Submitted 8 March, 2024;
originally announced March 2024.
-
Non-equilibrium pathways to emergent polar supertextures
Authors:
Vladimir A. Stoica,
Tiannan Yang,
Sujit Das,
Yue Cao,
Huaiyu Wang,
Yuya Kubota,
Cheng Dai,
Hari Padmanabhan,
Yusuke Sato,
Anudeep Mangu,
Quynh L. Nguyen,
Zhan Zhang,
Disha Talreja,
Marc E. Zajac,
Donald A. Walko,
Anthony D. DiChiara,
Shigeki Owada,
Kohei Miyanishi,
Kenji Tamasaku,
Takahiro Sato,
James M. Glownia,
Vincent Esposito,
Silke Nelson,
Matthias C. Hoffmann,
Richard D. Schaller
, et al. (9 additional authors not shown)
Abstract:
Ultrafast stimuli can stabilize metastable states of matter inaccessible by equilibrium means. Establishing the spatiotemporal link between ultrafast excitation and metastability is crucial to understanding these phenomena. Here, we use single-shot optical-pump, X-ray-probe measurements to provide snapshots of the emergence of a persistent polar vortex supercrystal in a heterostructure that hosts…
▽ More
Ultrafast stimuli can stabilize metastable states of matter inaccessible by equilibrium means. Establishing the spatiotemporal link between ultrafast excitation and metastability is crucial to understanding these phenomena. Here, we use single-shot optical-pump, X-ray-probe measurements to provide snapshots of the emergence of a persistent polar vortex supercrystal in a heterostructure that hosts a fine balance between built-in electrostatic and elastic frustrations by design. By perturbing this balance with photoinduced charges, a starting heterogenous mixture of polar phases disorders within a few picoseconds, resulting in a soup state composed of disordered ferroelectric and suppressed vortex orders. On the pico-to-nanosecond timescales, transient labyrinthine fluctuations form in this soup along with a recovering vortex order. On longer timescales, these fluctuations are progressively quenched by dynamical strain modulations, which drive the collective emergence of a single supercrystal phase. Our results, corroborated by dynamical phase-field modeling, reveal how ultrafast excitation of designer systems generates pathways for persistent metastability.
△ Less
Submitted 18 February, 2024;
originally announced February 2024.
-
Fine-tuning Reinforcement Learning Models is Secretly a Forgetting Mitigation Problem
Authors:
Maciej Wołczyk,
Bartłomiej Cupiał,
Mateusz Ostaszewski,
Michał Bortkiewicz,
Michał Zając,
Razvan Pascanu,
Łukasz Kuciński,
Piotr Miłoś
Abstract:
Fine-tuning is a widespread technique that allows practitioners to transfer pre-trained capabilities, as recently showcased by the successful applications of foundation models. However, fine-tuning reinforcement learning (RL) models remains a challenge. This work conceptualizes one specific cause of poor transfer, accentuated in the RL setting by the interplay between actions and observations: for…
▽ More
Fine-tuning is a widespread technique that allows practitioners to transfer pre-trained capabilities, as recently showcased by the successful applications of foundation models. However, fine-tuning reinforcement learning (RL) models remains a challenge. This work conceptualizes one specific cause of poor transfer, accentuated in the RL setting by the interplay between actions and observations: forgetting of pre-trained capabilities. Namely, a model deteriorates on the state subspace of the downstream task not visited in the initial phase of fine-tuning, on which the model behaved well due to pre-training. This way, we lose the anticipated transfer benefits. We identify conditions when this problem occurs, showing that it is common and, in many cases, catastrophic. Through a detailed empirical analysis of the challenging NetHack and Montezuma's Revenge environments, we show that standard knowledge retention techniques mitigate the problem and thus allow us to take full advantage of the pre-trained capabilities. In particular, in NetHack, we achieve a new state-of-the-art for neural models, improving the previous best score from $5$K to over $10$K points in the Human Monk scenario.
△ Less
Submitted 17 July, 2024; v1 submitted 5 February, 2024;
originally announced February 2024.
-
Exploiting Novel GPT-4 APIs
Authors:
Kellin Pelrine,
Mohammad Taufeeque,
Michał Zając,
Euan McLean,
Adam Gleave
Abstract:
Language model attacks typically assume one of two extreme threat models: full white-box access to model weights, or black-box access limited to a text generation API. However, real-world APIs are often more flexible than just text generation: these APIs expose "gray-box" access leading to new threat vectors. To explore this, we red-team three new functionalities exposed in the GPT-4 APIs: fine-tu…
▽ More
Language model attacks typically assume one of two extreme threat models: full white-box access to model weights, or black-box access limited to a text generation API. However, real-world APIs are often more flexible than just text generation: these APIs expose "gray-box" access leading to new threat vectors. To explore this, we red-team three new functionalities exposed in the GPT-4 APIs: fine-tuning, function calling and knowledge retrieval. We find that fine-tuning a model on as few as 15 harmful examples or 100 benign examples can remove core safeguards from GPT-4, enabling a range of harmful outputs. Furthermore, we find that GPT-4 Assistants readily divulge the function call schema and can be made to execute arbitrary function calls. Finally, we find that knowledge retrieval can be hijacked by injecting instructions into retrieval documents. These vulnerabilities highlight that any additions to the functionality exposed by an API can create new vulnerabilities.
△ Less
Submitted 4 August, 2024; v1 submitted 21 December, 2023;
originally announced December 2023.
-
Infrared Nanoimaging of Hydrogenated Perovskite Nickelate Synaptic Devices
Authors:
Sampath Gamage,
Sukriti Manna,
Marc Zajac,
Steven Hancock,
Qi Wang,
Sarabpreet Singh,
Mahdi Ghafariasl,
Kun Yao,
Tom Tiwald,
Tae Joon Park,
David P. Landau,
Haidan Wen,
Subramanian Sankaranarayanan,
Pierre Darancet,
Shriram Ramanathan,
Yohannes Abate
Abstract:
Solid-state devices made from correlated oxides such as perovskite nickelates are promising for neuromorphic computing by mimicking biological synaptic function. However, comprehending dopant action at the nanoscale poses a formidable challenge to understanding the elementary mechanisms involved. Here, we perform operando infrared nanoimaging of hydrogen-doped correlated perovskite, neodymium nick…
▽ More
Solid-state devices made from correlated oxides such as perovskite nickelates are promising for neuromorphic computing by mimicking biological synaptic function. However, comprehending dopant action at the nanoscale poses a formidable challenge to understanding the elementary mechanisms involved. Here, we perform operando infrared nanoimaging of hydrogen-doped correlated perovskite, neodymium nickel oxide (H-NdNiO3) devices and reveal how an applied field perturbs dopant distribution at the nanoscale. This perturbation leads to stripe phases of varying conductivity perpendicular to the applied field, which define the macroscale electrical characteristics of the devices. Hyperspectral nano-FTIR imaging in conjunction with density functional theory calculations unveil a real-space map of multiple vibrational states of H-NNO associated with OH stretching modes and their dependence on the dopant concentration. Moreover, the localization of excess charges induces an out-of-plane lattice expansion in NNO which was confirmed by in-situ - x-ray diffraction and creates a strain that acts as a barrier against further diffusion. Our results and the techniques presented here hold great potential to the rapidly growing field of memristors and neuromorphic devices wherein nanoscale ion motion is fundamentally responsible for function.
△ Less
Submitted 29 August, 2023;
originally announced September 2023.
-
Hamiltonian stochastic Lie systems and applications
Authors:
J. de Lucas,
X. Rivas,
M. Zajac
Abstract:
This paper provides a practical approach to stochastic Lie systems, i.e. stochastic differential equations whose general solutions can be written as a function depending only on a generic family of particular solutions and some constants, so as to emphasise their applications. We correct the known stochastic Lie theorem characterising stochastic Lie systems, proving that, contrary to previous clai…
▽ More
This paper provides a practical approach to stochastic Lie systems, i.e. stochastic differential equations whose general solutions can be written as a function depending only on a generic family of particular solutions and some constants, so as to emphasise their applications. We correct the known stochastic Lie theorem characterising stochastic Lie systems, proving that, contrary to previous claims, it satisfies the Malliavin's principle. Meanwhile, we show that stochastic Lie systems admit new stochastic features in the Ito approach. New generalisations of stochastic Lie systems, like the so-called stochastic foliated Lie systems, are devised. Subsequently, we focus on stochastic (foliated) Lie systems that can be studied as Hamiltonian systems using different types of differential geometric structures. We study their stability properties and we devise the basics of an energy-momentum method. A stochastic Poisson coalgebra method is developed to derive superposition rules for Hamiltonian stochastic Lie systems. Applications of our results are found in coronavirus stochastic models, stochastic Lotka-Volterra systems, stochastic SIS models of different types, etc. Our results improve previous approaches by using stochastic differential equations instead of deterministic models designed to grasp some of their stochastic features.
△ Less
Submitted 12 July, 2023;
originally announced July 2023.
-
Prediction Error-based Classification for Class-Incremental Learning
Authors:
Michał Zając,
Tinne Tuytelaars,
Gido M. van de Ven
Abstract:
Class-incremental learning (CIL) is a particularly challenging variant of continual learning, where the goal is to learn to discriminate between all classes presented in an incremental fashion. Existing approaches often suffer from excessive forgetting and imbalance of the scores assigned to classes that have not been seen together during training. In this study, we introduce a novel approach, Pre…
▽ More
Class-incremental learning (CIL) is a particularly challenging variant of continual learning, where the goal is to learn to discriminate between all classes presented in an incremental fashion. Existing approaches often suffer from excessive forgetting and imbalance of the scores assigned to classes that have not been seen together during training. In this study, we introduce a novel approach, Prediction Error-based Classification (PEC), which differs from traditional discriminative and generative classification paradigms. PEC computes a class score by measuring the prediction error of a model trained to replicate the outputs of a frozen random neural network on data from that class. The method can be interpreted as approximating a classification rule based on Gaussian Process posterior variance. PEC offers several practical advantages, including sample efficiency, ease of tuning, and effectiveness even when data are presented one class at a time. Our empirical results show that PEC performs strongly in single-pass-through-data CIL, outperforming other rehearsal-free baselines in all cases and rehearsal-based methods with moderate replay buffer size in most cases across multiple benchmarks.
△ Less
Submitted 9 March, 2024; v1 submitted 30 May, 2023;
originally announced May 2023.
-
Exploring Continual Learning of Diffusion Models
Authors:
Michał Zając,
Kamil Deja,
Anna Kuzina,
Jakub M. Tomczak,
Tomasz Trzciński,
Florian Shkurti,
Piotr Miłoś
Abstract:
Diffusion models have achieved remarkable success in generating high-quality images thanks to their novel training procedures applied to unprecedented amounts of data. However, training a diffusion model from scratch is computationally expensive. This highlights the need to investigate the possibility of training these models iteratively, reusing computation while the data distribution changes. In…
▽ More
Diffusion models have achieved remarkable success in generating high-quality images thanks to their novel training procedures applied to unprecedented amounts of data. However, training a diffusion model from scratch is computationally expensive. This highlights the need to investigate the possibility of training these models iteratively, reusing computation while the data distribution changes. In this study, we take the first step in this direction and evaluate the continual learning (CL) properties of diffusion models. We begin by benchmarking the most common CL methods applied to Denoising Diffusion Probabilistic Models (DDPMs), where we note the strong performance of the experience replay with the reduced rehearsal coefficient. Furthermore, we provide insights into the dynamics of forgetting, which exhibit diverse behavior across diffusion timesteps. We also uncover certain pitfalls of using the bits-per-dimension metric for evaluating CL.
△ Less
Submitted 27 March, 2023;
originally announced March 2023.
-
Characterizing non-Markovian Off-Resonant Errors in Quantum Gates
Authors:
Ken Xuan Wei,
Emily Pritchett,
David M. Zajac,
David C. McKay,
Seth Merkel
Abstract:
As quantum gates improve, it becomes increasingly difficult to characterize the remaining errors. Here we describe a class of coherent non-Markovian errors -- excitations due to an off-resonant drive -- that occur naturally in quantum devices that use time-dependent fields to generate gate operations. We show how these errors are mischaracterized using standard Quantum Computer Verification and Va…
▽ More
As quantum gates improve, it becomes increasingly difficult to characterize the remaining errors. Here we describe a class of coherent non-Markovian errors -- excitations due to an off-resonant drive -- that occur naturally in quantum devices that use time-dependent fields to generate gate operations. We show how these errors are mischaracterized using standard Quantum Computer Verification and Validation (QCVV) techniques that rely on Markovianity and are therefore often overlooked or assumed to be incoherent. We first demonstrate off-resonant errors within a simple toy model of Z-gates created by the AC Stark effect, then show how off-resonant errors manifest in all gates driven on a fixed-frequency transmon architecture, a prominent example being incidental cross-resonance interaction driven during single-qubit gates. Furthermore, the same methodology can access the errors caused by two-level systems (TLS), showing evidence of coherent, off-resonant interactions with subsystems that are not intentional qubits. While we explore these results and their impact on gate error for fixed-frequency devices, we note that off-resonant excitations potentially limit any architectures that use frequency selectivity.
△ Less
Submitted 8 February, 2024; v1 submitted 21 February, 2023;
originally announced February 2023.
-
Trust Your $\nabla$: Gradient-based Intervention Targeting for Causal Discovery
Authors:
Mateusz Olko,
Michał Zając,
Aleksandra Nowak,
Nino Scherrer,
Yashas Annadani,
Stefan Bauer,
Łukasz Kuciński,
Piotr Miłoś
Abstract:
Inferring causal structure from data is a challenging task of fundamental importance in science. Observational data are often insufficient to identify a system's causal structure uniquely. While conducting interventions (i.e., experiments) can improve the identifiability, such samples are usually challenging and expensive to obtain. Hence, experimental design approaches for causal discovery aim to…
▽ More
Inferring causal structure from data is a challenging task of fundamental importance in science. Observational data are often insufficient to identify a system's causal structure uniquely. While conducting interventions (i.e., experiments) can improve the identifiability, such samples are usually challenging and expensive to obtain. Hence, experimental design approaches for causal discovery aim to minimize the number of interventions by estimating the most informative intervention target. In this work, we propose a novel Gradient-based Intervention Targeting method, abbreviated GIT, that 'trusts' the gradient estimator of a gradient-based causal discovery framework to provide signals for the intervention acquisition function. We provide extensive experiments in simulated and real-world datasets and demonstrate that GIT performs on par with competitive baselines, surpassing them in the low-data regime.
△ Less
Submitted 3 April, 2024; v1 submitted 24 November, 2022;
originally announced November 2022.
-
Outan: An On-Head System for Driving micro-LED Arrays Implanted in Freely Moving Mice
Authors:
Alexander Tarnavsky Eitan,
Shirly Someck,
Mario Zajac,
Eran Socher,
Eran Stark
Abstract:
In the intact brain, neural activity can be recorded using sensing electrodes and manipulated using light stimulation. Silicon probes with integrated electrodes and micro-LEDs enable the detection and control of neural activity using a single implanted device. Miniaturized solutions for recordings from small freely moving animals are commercially available, but stimulation is driven by large, stat…
▽ More
In the intact brain, neural activity can be recorded using sensing electrodes and manipulated using light stimulation. Silicon probes with integrated electrodes and micro-LEDs enable the detection and control of neural activity using a single implanted device. Miniaturized solutions for recordings from small freely moving animals are commercially available, but stimulation is driven by large, stationary current sources. We designed and fabricated a current source chip and integrated it into a headstage PCB that weighs 1.37 g. The proposed system provides 10-bit resolution current control for 32 channels, driving micro-LEDs with up to 4.6 V and sourcing up to 0.9 mA at a refresh rate of 5 kHz per channel. When calibrated against a micro-LED probe, the system allows linear control of light output power, up to 10 micro-W per micro-LED. To demonstrate the capabilities of the system, synthetic sequences of neural spiking activity were produced by driving multiple micro-LEDs implanted in the hippocampal CA1 area of a freely moving mouse. The high spatial, temporal, and amplitude resolution of the system provides a rich variety of stimulation patterns. Combined with commercially available sampling headstages, the system provides an easy to use back-end, fully utilizing the bi-directional potential of integrated opto-electronic arrays.
△ Less
Submitted 24 November, 2022;
originally announced November 2022.
-
Dynamical criticality of spin-shear coupling in van der Waals antiferromagnets
Authors:
Faran Zhou,
Kyle Hwangbo,
Qi Zhang,
Chong Wang,
Lingnan Shen,
Jiawei Zhang,
Qianni Jiang,
Alfred Zong,
Yifan Su,
Marc Zajac,
Youngjun Ahn,
Donald Walko,
Richard Schaller,
Jiun-Haw Chu,
Nuh Gedik,
Xiaodong Xu,
Di Xiao,
Haidan Wen
Abstract:
The interplay between a multitude of electronic, spin, and lattice degrees of freedom underlies the complex phase diagrams of quantum materials. Layer stacking in van der Waals (vdW) heterostructures is responsible for exotic electronic and magnetic properties, which inspires stacking control of two-dimensional magnetism. Beyond the interplay between stacking order and interlayer magnetism, we dis…
▽ More
The interplay between a multitude of electronic, spin, and lattice degrees of freedom underlies the complex phase diagrams of quantum materials. Layer stacking in van der Waals (vdW) heterostructures is responsible for exotic electronic and magnetic properties, which inspires stacking control of two-dimensional magnetism. Beyond the interplay between stacking order and interlayer magnetism, we discover a spin-shear coupling mechanism in which a subtle shear of the atomic layers can have a profound effect on the intralayer magnetic order in a family of vdW antiferromagnets. Using time-resolved x-ray diffraction and optical linear dichroism measurements, interlayer shear is identified as the primary structural degree of freedom that couples with magnetic order. The recovery times of both shear and magnetic order upon optical excitation diverge at the magnetic ordering temperature with the same critical exponent. The time-dependent Ginzburg-Landau theory shows that this concurrent critical slowing down arises from a linear coupling of the interlayer shear to the magnetic order, which is dictated by the broken mirror symmetry intrinsic to the monoclinic stacking. Our results highlight the importance of interlayer shear in ultrafast control of magnetic order via spin-mechanical coupling.
△ Less
Submitted 15 November, 2022;
originally announced November 2022.
-
X-ray Free Electron Laser Studies of Electron and Phonon Dynamics of Graphene Adsorbed on Copper
Authors:
Hirohito Ogasawara,
Han Wang,
Jörgen Gladh,
Alessandro Gallo,
Ralph Page,
Johannes Voss,
Alan Luntz,
Elias Diesen,
Frank Abild-Pedersen,
Anders Nilsson,
Markus Soldemo,
Marc Zajac,
Andrew Attar,
Michelle E. Chen,
Sang Wan Cho,
Abhishek Katoch,
Ki-Jeong Kim,
Kyung Hwan Kim,
Minseok Kim,
Soonnam Kwon,
Sang Han Park,
Henrique Ribeiro,
Sami Sainio,
Hsin-Yi Wang,
Cheolhee Yang
, et al. (1 additional authors not shown)
Abstract:
We report optical pumping and X-ray absorption spectroscopy experiments at the PAL free electron laser that directly probe the electron dynamics of a graphene monolayer adsorbed on copper in the femtosecond regime. By analyzing the results with ab-initio theory we infer that the excitation of graphene is dominated by indirect excitation from hot electron-hole pairs created in the copper by the opt…
▽ More
We report optical pumping and X-ray absorption spectroscopy experiments at the PAL free electron laser that directly probe the electron dynamics of a graphene monolayer adsorbed on copper in the femtosecond regime. By analyzing the results with ab-initio theory we infer that the excitation of graphene is dominated by indirect excitation from hot electron-hole pairs created in the copper by the optical laser pulse. However, once the excitation is created in graphene, its decay follows a similar path as in many previous studies of graphene adsorbed on semiconductors, i e. rapid excitation of SCOPS (Strongly Coupled Optical Phonons) and eventual thermalization. It is likely that the lifetime of the hot electron-hole pairs in copper governs the lifetime of the electronic excitation of the graphene.
△ Less
Submitted 1 November, 2022;
originally announced November 2022.
-
Disentangling Transfer in Continual Reinforcement Learning
Authors:
Maciej Wołczyk,
Michał Zając,
Razvan Pascanu,
Łukasz Kuciński,
Piotr Miłoś
Abstract:
The ability of continual learning systems to transfer knowledge from previously seen tasks in order to maximize performance on new tasks is a significant challenge for the field, limiting the applicability of continual learning solutions to realistic scenarios. Consequently, this study aims to broaden our understanding of transfer and its driving forces in the specific case of continual reinforcem…
▽ More
The ability of continual learning systems to transfer knowledge from previously seen tasks in order to maximize performance on new tasks is a significant challenge for the field, limiting the applicability of continual learning solutions to realistic scenarios. Consequently, this study aims to broaden our understanding of transfer and its driving forces in the specific case of continual reinforcement learning. We adopt SAC as the underlying RL algorithm and Continual World as a suite of continuous control tasks. We systematically study how different components of SAC (the actor and the critic, exploration, and data) affect transfer efficacy, and we provide recommendations regarding various modeling options. The best set of choices, dubbed ClonEx-SAC, is evaluated on the recent Continual World benchmark. ClonEx-SAC achieves 87% final success rate compared to 80% of PackNet, the best method in the benchmark. Moreover, the transfer grows from 0.18 to 0.54 according to the metric provided by Continual World.
△ Less
Submitted 28 September, 2022;
originally announced September 2022.
-
A Discrete Hamilton--Jacobi Theory for Contact Hamiltonian Dynamics
Authors:
Oğul Esen,
Cristina Sardón,
Marcin Zajac
Abstract:
In this paper, we propose a discrete Hamilton--Jacobi theory for (discrete) Hamiltonian dynamics defined on a (discrete) contact manifold. To this end, we first provide a novel geometric Hamilton--Jacobi theory for continuous contact Hamiltonian dynamics. Then, rooting on the discrete contact Lagrangian formulation, we obtain the discrete equations for Hamiltonian dynamics by the discrete Legendre…
▽ More
In this paper, we propose a discrete Hamilton--Jacobi theory for (discrete) Hamiltonian dynamics defined on a (discrete) contact manifold. To this end, we first provide a novel geometric Hamilton--Jacobi theory for continuous contact Hamiltonian dynamics. Then, rooting on the discrete contact Lagrangian formulation, we obtain the discrete equations for Hamiltonian dynamics by the discrete Legendre transformation. Based on the discrete contact Hamilton equation, we construct a discrete Hamilton--Jacobi equation for contact Hamiltonian dynamics. We show how the discrete Hamilton--Jacobi equation is related to the continuous Hamilton--Jacobi theory presented in this work. Then, we propose geometric foundations of the discrete Hamilton--Jacobi equations on contact manifolds in terms of discrete contact flows. At the end of the paper we provide a numerical example to test the theory.
△ Less
Submitted 13 September, 2022;
originally announced September 2022.
-
Mrs. Correct and Majority Colorings
Authors:
Marcin Anholcer,
Bartłomiej Bosek,
Jarosław Grytczuk,
Grzegorz Gutowski,
Jakub Przybyło,
Mariusz Zając
Abstract:
A majority coloring of a directed graph is a vertex coloring in which each vertex has the same color as at most half of its out-neighbors. In this note we simplify some proof techniques and generalize previously known results on various generalizations of majority coloring. In particular, our unified and simplified approach works for paintability - an on-line analog of the list coloring.
A majority coloring of a directed graph is a vertex coloring in which each vertex has the same color as at most half of its out-neighbors. In this note we simplify some proof techniques and generalize previously known results on various generalizations of majority coloring. In particular, our unified and simplified approach works for paintability - an on-line analog of the list coloring.
△ Less
Submitted 5 September, 2022; v1 submitted 20 July, 2022;
originally announced July 2022.
-
The Tulczyjew triple for gauge field theories
Authors:
Marcin Zając
Abstract:
In this article we construct and discuss a new rigorous geometric formalism for gauge field theories. The basis of our work is the notion of the Tulczyjew triple, a geometric structure which successfully solved numerous problems in mathematical description of mechanics and classical field theory. In particular, we construct a Tulczyjew triple for gauge field theories and reduce it for systems that…
▽ More
In this article we construct and discuss a new rigorous geometric formalism for gauge field theories. The basis of our work is the notion of the Tulczyjew triple, a geometric structure which successfully solved numerous problems in mathematical description of mechanics and classical field theory. In particular, we construct a Tulczyjew triple for gauge field theories and reduce it for systems that depend only on the value of a connection and curvature instead of the entire first jet of the gauge field. We also introduce new geometric structures such as the vector-affine product of bundles and analyse the connection bundle from a new perspective. Finally, we apply the derived formalism to Yang-Mills theory.
△ Less
Submitted 1 April, 2022;
originally announced April 2022.
-
Phonon-Mediated Attractive Interactions between Excitons in Lead-Halide-Perovskites
Authors:
Nuri Yazdani,
Maryna I. Bodnarchuk,
Federica Bertolotti,
Norberto Masciocchi,
Ina Fureraj,
Burak Guzelturk,
Benjamin L. Cotts,
Marc Zajac,
Gabriele Rainò,
Maximilian Jansen,
Simon C. Boehme,
Maksym Yarema,
Ming-Fu Lin,
Michael Kozina,
Alexander Reid,
Xiaozhe Shen,
Stephen Weathersby,
Xijie Wang,
Eric Vauthey,
Antonietta Guagliardi,
Maksym V. Kovalenko,
Vanessa Wood,
Aaron Lindenberg
Abstract:
Understanding the origin of electron-phonon coupling in lead-halide perovskites (LHP) is key to interpreting and leveraging their optical and electronic properties. Here we perform femtosecond-resolved, optical-pump, electron-diffraction-probe measurements to quantify the lattice reorganization occurring as a result of photoexcitation in LHP nanocrystals. Photoexcitation is found to drive a reduct…
▽ More
Understanding the origin of electron-phonon coupling in lead-halide perovskites (LHP) is key to interpreting and leveraging their optical and electronic properties. Here we perform femtosecond-resolved, optical-pump, electron-diffraction-probe measurements to quantify the lattice reorganization occurring as a result of photoexcitation in LHP nanocrystals. Photoexcitation is found to drive a reduction in lead-halide octahedra tilts and distortions in the lattice, a result of deformation potential coupling to low energy optical phonons. Our results indicate particularly strong coupling in FAPbBr3, and far weaker coupling in CsPbBr3, highlighting differences in the dominant machanisms governing electron-phonon coupling in LHPs. We attribute the enhanced coupling in FAPbBr3 to its disordered crystal structure, which persists down to cryogenic temperatures. We find the reorganizations induced by each exciton in a multiexcitonic state constructively interfere, giving rise to a coupling strength which scales quadratically with the exciton number. This superlinear scaling induces phonon-mediated attractive interactions between excitations in LHPs.
△ Less
Submitted 11 March, 2022;
originally announced March 2022.
-
Reviewing the Geometric Hamilton-Jacobi Theory concerning Jacobi and Leibniz identities
Authors:
O. Esen,
M. de León,
M. Lainz,
C. Sardón,
M. Zając
Abstract:
In this survey, we review the classical Hamilton Jacobi theory from a geometric point of view in different geometric backgrounds. We propose a Hamilton Jacobi equation for different geometric structures attending to one particular characterization: whether they fulfill the Jacobi and Leibniz identities simultaneously, or if at least they satisfy one of them.
In this regard, we review the case of…
▽ More
In this survey, we review the classical Hamilton Jacobi theory from a geometric point of view in different geometric backgrounds. We propose a Hamilton Jacobi equation for different geometric structures attending to one particular characterization: whether they fulfill the Jacobi and Leibniz identities simultaneously, or if at least they satisfy one of them.
In this regard, we review the case of time dependent and dissipative physical systems as systems that fulfill the Jacobi identity but not the Leibnitz identity. Furthermore, we review the contact evolution Hamilton Jacobi theory as a split off the regular contact geometry, and that actually satisfies the Leibniz rule instead of Jacobi.
Furthermore, we include a novel result, which is the Hamilton-Jacobi equation for conformal Hamiltonian vector fields as a generalization of the well known Hamilton Jacobi on a symplectic manifold, that is retrieved in the case of a zero conformal factor. The interest of a geometric Hamilton Jacobi equation is the primordial observation that if a Hamiltonian vector field can be projected into a configuration manifold by means of a 1-form dW, then the integral curves of the projected vector field can be transformed into integral curves of the Hamiltonian vector field provided that W is a solution of the Hamilton-Jacobi equation. Geometrically, the solution of the Hamilton Jacobi equation plays the role of a Lagrangian submanifold of a certain bundle. Exploiting these features in different geometric scenarios we propose a geometric theory for multiple physical systems depending on the fundamental identities that their dynamic satisfies. Different examples are pictured to reflect the results provided, being all of them new, except for one that is reassessment of a previously considered example.
△ Less
Submitted 14 February, 2022;
originally announced February 2022.
-
On a Problem of Steinhaus
Authors:
Marcin Anholcer,
Bartłomiej Bosek,
Jarosław Grytczuk,
Grzegorz Gutowski,
Jakub Przybyło,
Rafał Pyzik,
Mariusz Zając
Abstract:
Let $N$ be a positive integer. A sequence $X=(x_1,x_2,\ldots,x_N)$ of points in the unit interval $[0,1)$ is piercing if $\{x_1,x_2,\ldots,x_n\}\cap \left[\frac{i}{n},\frac{i+1}{n} \right) \neq\emptyset$ holds for every $n=1,2,\ldots, N$ and every $i=0,1,\ldots,n-1$. In 1958 Steinhaus asked whether piercing sequences can be arbitrarily long. A negative answer was provided by Schinzel, who proved t…
▽ More
Let $N$ be a positive integer. A sequence $X=(x_1,x_2,\ldots,x_N)$ of points in the unit interval $[0,1)$ is piercing if $\{x_1,x_2,\ldots,x_n\}\cap \left[\frac{i}{n},\frac{i+1}{n} \right) \neq\emptyset$ holds for every $n=1,2,\ldots, N$ and every $i=0,1,\ldots,n-1$. In 1958 Steinhaus asked whether piercing sequences can be arbitrarily long. A negative answer was provided by Schinzel, who proved that any such sequence may have at most $74$ elements. This was later improved to the best possible value of $17$ by Warmus, and independently by Berlekamp and Graham.
In this paper we study a more general variant of piercing sequences. Let $f(n)\geq n$ be an infinite nondecreasing sequence of positive integers. A sequence $X=(x_1,x_2,\ldots,x_{f(N)})$ is $f$-piercing if $\{x_1,x_2,\ldots,x_{f(n)}\}\cap \left[\frac{i}{n},\frac{i+1}{n} \right) \neq\emptyset$ holds for every $n=1,2,\ldots, N$ and every $i=0,1,\ldots,n-1$. A special case of $f(n)=n+d$, with $d$ a fixed nonnegative integer, was studied by Berlekamp and Graham. They noticed that for each $d\geq 0$, the maximum length of any $(n+d)$-piercing sequence is finite. Expressing this maximum length as $s(d)+d$, they obtained an exponential upper bound on the function $s(d)$, which was later improved to $s(d)=O(d^3)$ by Graham and Levy. Recently, Konyagin proved that $2d\leqslant s(d)< 200d$ holds for all sufficiently big $d$.
Using a different technique based on the Farey fractions and stick-breaking games, we prove here that the function $s(d)$ satisfies $\left\lfloor{}c_1d\right\rfloor{}\leqslant s(d)\leqslant c_2d+o(d)$, where $c_1=\frac{\ln 2}{1-\ln 2}\approx2.25$ and $c_2=\frac{1+\ln2}{1-\ln2}\approx5.52$. We also prove that there exists an infinite $f$-piercing sequence with $f(n)= γn+o(n)$ if and only if $γ\geq\frac{1}{\ln 2}\approx 1.44$.
△ Less
Submitted 15 December, 2021; v1 submitted 2 November, 2021;
originally announced November 2021.
-
Spectator Errors in Tunable Coupling Architectures
Authors:
D. M. Zajac,
J. Stehlik,
D. L. Underwood,
T. Phung,
J. Blair,
S. Carnevale,
D. Klaus,
G. A. Keefe,
A. Carniol,
M. Kumph,
Matthias Steffen,
O. E. Dial
Abstract:
The addition of tunable couplers to superconducting quantum architectures offers significant advantages for scaling compared to fixed coupling approaches. In principle, tunable couplers allow for exact cancellation of qubit-qubit coupling through the interference of two parallel coupling pathways between qubits. However, stray microwave couplings can introduce additional pathways which complicate…
▽ More
The addition of tunable couplers to superconducting quantum architectures offers significant advantages for scaling compared to fixed coupling approaches. In principle, tunable couplers allow for exact cancellation of qubit-qubit coupling through the interference of two parallel coupling pathways between qubits. However, stray microwave couplings can introduce additional pathways which complicate the interference effect. Here we investigate the primary spectator induced errors of the bus below qubit (BBQ) architecture in a six qubit device. We identify the key design parameters which inhibit ideal cancellation and demonstrate that dynamic cancellation pulses can further mitigate spectator errors.
△ Less
Submitted 25 August, 2021;
originally announced August 2021.
-
Microscopic processes during ultra-fast laser generation of Frenkel defects in diamond
Authors:
Benjamin Griffiths,
Andrew Kirkpatrick,
Shannon S. Nicley,
Rajesh L. Patel,
Joanna M. Zajac,
Gavin W. Morley,
Martin J. Booth,
Patrick S. Salter,
Jason M. Smith
Abstract:
Engineering single atomic defects into wide bandgap materials has become an attractive field in recent years due to emerging applications such as solid-state quantum bits and sensors. The simplest atomic-scale defect is the lattice vacancy which is often a constituent part of more complex defects such as the nitrogen-vacancy (NV) centre in diamond, therefore an understanding of the formation mecha…
▽ More
Engineering single atomic defects into wide bandgap materials has become an attractive field in recent years due to emerging applications such as solid-state quantum bits and sensors. The simplest atomic-scale defect is the lattice vacancy which is often a constituent part of more complex defects such as the nitrogen-vacancy (NV) centre in diamond, therefore an understanding of the formation mechanisms and precision engineering of vacancies is desirable. We present a theoretical and experimental study into the ultra-fast laser generation of vacancy-interstitial pairs (Frenkel defects) in diamond. The process is described by a set of coupled rate equations of the pulsed laser interaction with the material and of the non-equilibrium dynamics of charge carriers during and in the wake of the pulse. We find that a model for Frenkel defect generation via the recombination of a bound biexciton as the electron plasma cools provides good agreement with experimental data, reproducing an effective non-linearity of $\sim$ 40 for Frenkel defect generation with respect to laser pulse energy.
△ Less
Submitted 25 May, 2021;
originally announced May 2021.
-
Continual World: A Robotic Benchmark For Continual Reinforcement Learning
Authors:
Maciej Wołczyk,
Michał Zając,
Razvan Pascanu,
Łukasz Kuciński,
Piotr Miłoś
Abstract:
Continual learning (CL) -- the ability to continuously learn, building on previously acquired knowledge -- is a natural requirement for long-lived autonomous reinforcement learning (RL) agents. While building such agents, one needs to balance opposing desiderata, such as constraints on capacity and compute, the ability to not catastrophically forget, and to exhibit positive transfer on new tasks.…
▽ More
Continual learning (CL) -- the ability to continuously learn, building on previously acquired knowledge -- is a natural requirement for long-lived autonomous reinforcement learning (RL) agents. While building such agents, one needs to balance opposing desiderata, such as constraints on capacity and compute, the ability to not catastrophically forget, and to exhibit positive transfer on new tasks. Understanding the right trade-off is conceptually and computationally challenging, which we argue has led the community to overly focus on catastrophic forgetting. In response to these issues, we advocate for the need to prioritize forward transfer and propose Continual World, a benchmark consisting of realistic and meaningfully diverse robotic tasks built on top of Meta-World as a testbed. Following an in-depth empirical evaluation of existing CL methods, we pinpoint their limitations and highlight unique algorithmic challenges in the RL setting. Our benchmark aims to provide a meaningful and computationally inexpensive challenge for the community and thus help better understand the performance of existing and future solutions. Information about the benchmark, including the open-source code, is available at https://sites.google.com/view/continualworld.
△ Less
Submitted 28 October, 2021; v1 submitted 23 May, 2021;
originally announced May 2021.
-
The dressing field method in gauge theories -- geometric approach
Authors:
Marcin Zając
Abstract:
Recently, T. Masson, J. Francois, S. Lazzarini, C. Fournel and J. Attard have introduced a new method of the reduction of gauge symmetry called a dressing field method. In this paper we analyse this method from the fiber bundle point of view and we show the geometric implications for a principal bundle underlying a given gauge theory. We show how the existence of the dressing field satisfying cert…
▽ More
Recently, T. Masson, J. Francois, S. Lazzarini, C. Fournel and J. Attard have introduced a new method of the reduction of gauge symmetry called a dressing field method. In this paper we analyse this method from the fiber bundle point of view and we show the geometric implications for a principal bundle underlying a given gauge theory. We show how the existence of the dressing field satisfying certain conditions naturally leads to the reduction of the principal bundle and, as a consequence, to the reduction of the configuration and phase bundle of the system.
△ Less
Submitted 20 May, 2021;
originally announced May 2021.
-
Time-dependent Hamiltonian mechanics on a locally conformal symplectic manifold
Authors:
Orlando Ragnisco,
Cristina Sardon,
Marcin Zając
Abstract:
In this paper we aim at presenting a concise but also comprehensive study of time-dependent (tdependent) Hamiltonian dynamics on a locally conformal symplectic (lcs) manifold. We present a generalized geometric theory of canonical transformations and formulate a time-dependent geometric Hamilton-Jacobi theory on lcs manifolds. In contrast to previous papers concerning locally conformal symplectic…
▽ More
In this paper we aim at presenting a concise but also comprehensive study of time-dependent (tdependent) Hamiltonian dynamics on a locally conformal symplectic (lcs) manifold. We present a generalized geometric theory of canonical transformations and formulate a time-dependent geometric Hamilton-Jacobi theory on lcs manifolds. In contrast to previous papers concerning locally conformal symplectic manifolds, here the introduction of the time dependency brings out interesting geometric properties, as it is the introduction of contact geometry in locally symplectic patches. To conclude, we show examples of the applications of our formalism, in particular, we present systems of differential equations with time-dependent parameters, which admit different physical interpretations as we shall point out.
△ Less
Submitted 6 April, 2021;
originally announced April 2021.
-
Matched Pair Analysis of Euler-Poincaré Flow on Hamiltonian Vector Fields
Authors:
Oğul Esen,
Cristina Sardón,
Marcin Zajac
Abstract:
In this paper we provide a matched pair decomposition of the space of symmetric contravariant tensors $\mathfrak{T}\mathcal{Q}$. From this procedure two complementary Lie subalgebras of $\mathfrak{T}\mathcal{Q}$ under \textit{mutual} interaction arise. Introducing a lift operator, the matched pair decomposition of the space of Hamiltonian vector fields is determined. According to these realization…
▽ More
In this paper we provide a matched pair decomposition of the space of symmetric contravariant tensors $\mathfrak{T}\mathcal{Q}$. From this procedure two complementary Lie subalgebras of $\mathfrak{T}\mathcal{Q}$ under \textit{mutual} interaction arise. Introducing a lift operator, the matched pair decomposition of the space of Hamiltonian vector fields is determined. According to these realizations, Euler-Poincaré flows on such spaces are decomposed into two subdynamics: one of which is the Euler--Poincaré formulation of isentropic fluid flows, and the other one corresponds with Euler--Poincaré equations on higher order contravariant tensors ($n\geq 2$)
△ Less
Submitted 7 March, 2021;
originally announced March 2021.
-
Universal phase dynamics in VO2 switches revealed by ultrafast operando diffraction
Authors:
Aditya Sood,
Xiaozhe Shen,
Yin Shi,
Suhas Kumar,
Su Ji Park,
Marc Zajac,
Yifei Sun,
Long-Qing Chen,
Shriram Ramanathan,
Xijie Wang,
William C. Chueh,
Aaron M. Lindenberg
Abstract:
Strongly correlated materials that exhibit an insulator-metal transition are key candidates in the search for new computing platforms. Understanding the pathways and timescales underlying the electrically-driven insulator-metal transition is crucial for uncovering the fundamental limits of device operation. Using stroboscopic electron diffraction, we perform synchronized time-resolved measurements…
▽ More
Strongly correlated materials that exhibit an insulator-metal transition are key candidates in the search for new computing platforms. Understanding the pathways and timescales underlying the electrically-driven insulator-metal transition is crucial for uncovering the fundamental limits of device operation. Using stroboscopic electron diffraction, we perform synchronized time-resolved measurements of atomic motions and electronic transport in operating vanadium dioxide switches. We discover an electrically-triggered, isostructural state that forms transiently on microsecond timescales, stabilized by local heterogeneities and interfacial interactions between the equilibrium phases. This metastable phase bears striking similarity to that formed under photoexcitation within picoseconds, suggesting a universal transformation pathway across eight orders of magnitude of timescale. Our results establish a new route for uncovering non-equilibrium and metastable phases in correlated materials, and open avenues for engineering novel dynamical behavior in nanoelectronics.
△ Less
Submitted 11 February, 2021;
originally announced February 2021.
-
Tunable Coupling Architecture for Fixed-frequency Transmons
Authors:
J. Stehlik,
D. M. Zajac,
D. L. Underwood,
T. Phung,
J. Blair,
S. Carnevale,
D. Klaus,
G. A. Keefe,
A. Carniol,
M. Kumph,
Matthias Steffen,
O. E. Dial
Abstract:
Implementation of high-fidelity two-qubit operations is a key ingredient for scalable quantum error correction. In superconducting qubit architectures tunable buses have been explored as a means to higher fidelity gates. However, these buses introduce new pathways for leakage. Here we present a modified tunable bus architecture appropriate for fixed-frequency qubits in which the adiabaticity restr…
▽ More
Implementation of high-fidelity two-qubit operations is a key ingredient for scalable quantum error correction. In superconducting qubit architectures tunable buses have been explored as a means to higher fidelity gates. However, these buses introduce new pathways for leakage. Here we present a modified tunable bus architecture appropriate for fixed-frequency qubits in which the adiabaticity restrictions on gate speed are reduced. We characterize this coupler on a range of two-qubit devices achieving a maximum gate fidelity of $99.85\%$. We further show the calibration is stable over one day.
△ Less
Submitted 19 January, 2021;
originally announced January 2021.
-
Dynamic lattice distortions driven by surface trapping in semiconductor nanocrystals
Authors:
Burak Guzelturk,
Benjamin L. Cotts,
Dipti Jasrasaria,
John P. Philbin,
David A. Hanifi,
Brent A. Koscher,
Arunima D. Balan,
Ethan Curling,
Marc Zajac,
Suji Park,
Nuri Yazdani,
Clara Nyby,
Vladislav Kamysbayev,
Stefan Fischer,
Zach Nett,
Xiaozhe Shen,
Michael E. Kozina,
Ming-Fu Lin,
Alexander H. Reid,
Stephen P. Weathersby,
Richard D. Schaller,
Vanessa Wood,
Xijie Wang,
Jennifer A. Dionne,
Dmitri V. Talapin
, et al. (4 additional authors not shown)
Abstract:
Nonradiative processes limit optoelectronic functionality of nanocrystals and curb their device performance. Nevertheless, the dynamic structural origins of nonradiative relaxations in nanocrystals are not understood. Here, femtosecond electron diffraction measurements corroborated by atomistic simulations uncover transient lattice deformations accompanying radiationless electronic processes in se…
▽ More
Nonradiative processes limit optoelectronic functionality of nanocrystals and curb their device performance. Nevertheless, the dynamic structural origins of nonradiative relaxations in nanocrystals are not understood. Here, femtosecond electron diffraction measurements corroborated by atomistic simulations uncover transient lattice deformations accompanying radiationless electronic processes in semiconductor nanocrystals. Investigation of the excitation energy dependence shows that hot carriers created by a photon energy considerably larger than the bandgap induce structural distortions at nanocrystal surfaces on few picosecond timescales associated with the localization of trapped holes. On the other hand, carriers created by a photon energy close to the bandgap result in transient lattice heating that occurs on a much longer 200 ps timescale, governed by an Auger heating mechanism. Elucidation of the structural deformations associated with the surface trapping of hot holes provides atomic-scale insights into the mechanisms deteriorating optoelectronic performance and a pathway towards minimizing these losses in nanocrystal devices.
△ Less
Submitted 5 January, 2021;
originally announced January 2021.
-
The Tierras Observatory: An ultra-precise photometer to characterize nearby terrestrial exoplanets
Authors:
Juliana García-Mejía,
David Charbonneau,
Daniel Fabricant,
Jonathan M. Irwin,
Robert Fata,
Joseph M. Zajac,
Peter E. Doherty
Abstract:
We report on the status of the Tierras Observatory, a refurbished 1.3-m ultra-precise fully-automated photometer located at the F. L. Whipple Observatory atop Mt. Hopkins, Arizona. Tierras is designed to limit systematic errors, notably precipitable water vapor (PWV), to 250 ppm, enabling the characterization of terrestrial planet transits orbiting $< 0.3 \, R_{\odot}$ stars, as well as the potent…
▽ More
We report on the status of the Tierras Observatory, a refurbished 1.3-m ultra-precise fully-automated photometer located at the F. L. Whipple Observatory atop Mt. Hopkins, Arizona. Tierras is designed to limit systematic errors, notably precipitable water vapor (PWV), to 250 ppm, enabling the characterization of terrestrial planet transits orbiting $< 0.3 \, R_{\odot}$ stars, as well as the potential discovery of exo-moons and exo-rings. The design choices that will enable our science goals include: a four-lens focal reducer and field-flattener to increase the field-of-view of the telescope from a $11.94'$ to a $0.48^{\circ}$ side; a custom narrow bandpass ($40.2$ nm FWHM) filter centered around $863.5$ nm to minimize PWV errors known to limit ground-based photometry of red dwarfs; and a deep-depletion $4K \times 4K$ CCD with a 300ke-full well and QE$>85\%$ in our bandpass, operating in frame transfer mode. We are also pursuing the design of a set of baffles to minimize the significant amount of scattered light reaching the image plane. Tierras will begin science operations in early 2021.
△ Less
Submitted 17 December, 2020;
originally announced December 2020.
-
Graph Polynomials and Group Coloring of Graphs
Authors:
Bartłomiej Bosek,
Jarosław Grytczuk,
Grzegorz Gutowski,
Oriol Serra,
Mariusz Zając
Abstract:
Let $Γ$ be an Abelian group and let $G$ be a simple graph. We say that $G$ is $Γ$-colorable if for some fixed orientation of $G$ and every edge labeling $\ell:E(G)\rightarrow Γ$, there exists a vertex coloring $c$ by the elements of $Γ$ such that $c(y)-c(x)\neq \ell(e)$, for every edge $e=xy$ (oriented from $x$ to $y$).
Langhede and Thomassen proved recently that every planar graph on $n$ vertic…
▽ More
Let $Γ$ be an Abelian group and let $G$ be a simple graph. We say that $G$ is $Γ$-colorable if for some fixed orientation of $G$ and every edge labeling $\ell:E(G)\rightarrow Γ$, there exists a vertex coloring $c$ by the elements of $Γ$ such that $c(y)-c(x)\neq \ell(e)$, for every edge $e=xy$ (oriented from $x$ to $y$).
Langhede and Thomassen proved recently that every planar graph on $n$ vertices has at least $2^{n/9}$ different $\mathbb{Z}_5$-colorings. By using a different approach based on graph polynomials, we extend this result to $K_5$-minor-free graphs in the more general setting of field coloring. More specifically, we prove that every such graph on $n$ vertices is $\mathbb{F}$-$5$-choosable, whenever $\mathbb{F}$ is an arbitrary field with at least $5$ elements. Moreover, the number of colorings (for every list assignment) is at least $5^{n/4}$.
△ Less
Submitted 15 December, 2021; v1 submitted 6 December, 2020;
originally announced December 2020.
-
Geometry and solutions of an epidemic SIS model permitting fluctuations and quantization
Authors:
Oğul Esen,
Eduardo Fernández-Saiz,
Cristina Sardón,
Marcin Zając
Abstract:
Some recent works reveal that there are models of differential equations for the mean and variance of infected individuals that reproduce the SIS epidemic model at some point. This stochastic SIS epidemic model can be interpreted as a Hamiltonian system, therefore we wondered if it could be geometrically handled through the theory of Lie--Hamilton systems, and this happened to be the case. The pri…
▽ More
Some recent works reveal that there are models of differential equations for the mean and variance of infected individuals that reproduce the SIS epidemic model at some point. This stochastic SIS epidemic model can be interpreted as a Hamiltonian system, therefore we wondered if it could be geometrically handled through the theory of Lie--Hamilton systems, and this happened to be the case. The primordial result is that we are able to obtain a general solution for the stochastic/ SIS-epidemic model (with fluctuations) in form of a nonlinear superposition rule that includes particular stochastic solutions and certain constants to be related to initial conditions of the contagion process. The choice of these initial conditions will be crucial to display the expected behavior of the curve of infections during the epidemic. We shall limit these constants to nonsingular regimes and display graphics of the behavior of the solutions. As one could expect, the increase of infected individuals follows a sigmoid-like curve.
Lie--Hamiltonian systems admit a quantum deformation, so does the stochastic SIS-epidemic model. We present this generalization as well. If one wants to study the evolution of an SIS epidemic under the influence of a constant heat source (like centrally heated buildings), one can make use of quantum stochastic differential equations coming from the so-called quantum deformation.
△ Less
Submitted 6 August, 2020;
originally announced August 2020.
-
Graph polynomials and paintability of plane graphs
Authors:
Jarosław Grytczuk,
Stanislav Jendrol',
Mariusz Zając
Abstract:
There exists a variety of coloring problems for plane graphs, involving vertices, edges, and faces in all possible combinations. For instance, in the \emph{entire coloring} of a plane graph we are to color these three sets so that any pair of adjacent or incident elements get different colors. We study here some problems of this type from algebraic perspective, focusing on the \emph{facial} varian…
▽ More
There exists a variety of coloring problems for plane graphs, involving vertices, edges, and faces in all possible combinations. For instance, in the \emph{entire coloring} of a plane graph we are to color these three sets so that any pair of adjacent or incident elements get different colors. We study here some problems of this type from algebraic perspective, focusing on the \emph{facial} variant. We obtain several results concerning the \emph{Alon-Tarsi number} of various graphs derived from plane embeddings. This allows for extensions of some previous results for \emph{choosability} of these graphs to the game theoretic variant, know as \emph{paintability}. For instance, we prove that every plane graph is facially entirely \emph{$8$-paintable}, which means (metaphorically) that even a color-blind person can facially color the entire graph form lists of size $8$.
△ Less
Submitted 5 April, 2020;
originally announced April 2020.
-
Cauchy data space and multisymplectic formulation of conformal classical field theories
Authors:
Ogul Esen,
Manuel de León,
Cristina Sardón,
Marcin Zając
Abstract:
In this article we inspect the dynamics of classical field theories with a local conformal behavior. Our interest in the multisymplectic setting comes from its suitable description of field theories, and the conformal character has been added to account for field theories that are scale invariant, flat spaces, and because some conformal fields can be exactly solved or classified. In particular, we…
▽ More
In this article we inspect the dynamics of classical field theories with a local conformal behavior. Our interest in the multisymplectic setting comes from its suitable description of field theories, and the conformal character has been added to account for field theories that are scale invariant, flat spaces, and because some conformal fields can be exactly solved or classified. In particular, we will solve the example of a conformal scalar field using the geometric Hamilton-Jacobi theory that is explicitly proposed for conformal fields on a multisymplectic manifold. To complete the geometric approach to study field theories, we propose the Hamilton-Jacobi theory for conformal fields in a Cauchy data space, in which space and time are split separately and the dynamics is depicted in an infinite-dimensional manifold.
△ Less
Submitted 10 February, 2020;
originally announced February 2020.
-
Laurent polynomials of planar graphs
Authors:
Mariusz Zając
Abstract:
By using Laurent graph polynomials instead of the usual ones, i.e. by allowing negative powers of the variables, we simplify an existing method of determining the Alon-Tarsi numbers of planar graphs.
By using Laurent graph polynomials instead of the usual ones, i.e. by allowing negative powers of the variables, we simplify an existing method of determining the Alon-Tarsi numbers of planar graphs.
△ Less
Submitted 6 December, 2019;
originally announced December 2019.
-
The globalization problem of the Hamilton-DeDonder-Weyl equations on a local $k$-symplectic framework
Authors:
Oğul Esen,
Manuel de León,
Cristina Sardón,
Marcin Zajac
Abstract:
In this paper we aim at addressing the globalization problem of Hamilton-DeDonder-Weyl equations on a local $k$-symplectic framework and we introduce the notion of {\it locally conformal $k$-symplectic (l.c.k-s.) manifolds}. This formalism describes the dynamical properties of physical systems that locally behave like multi-Hamiltonian systems. Here, we describe the local Hamiltonian properties of…
▽ More
In this paper we aim at addressing the globalization problem of Hamilton-DeDonder-Weyl equations on a local $k$-symplectic framework and we introduce the notion of {\it locally conformal $k$-symplectic (l.c.k-s.) manifolds}. This formalism describes the dynamical properties of physical systems that locally behave like multi-Hamiltonian systems. Here, we describe the local Hamiltonian properties of such systems, but we also provide a global outlook by introducing the global Lee one-form approach. In particular, the dynamics will be depicted with the aid of the Hamilton--Jacobi equation, which is specifically proposed in a l.c.k-s manifold.
△ Less
Submitted 14 November, 2019;
originally announced November 2019.
-
Hamilton-Jacobi Formalism on Locally Conformally Symplectic Manifolds
Authors:
Oğul Esen,
Manuel de León,
Cristina Sardón,
Marcin Zajac
Abstract:
In this article we provide a Hamilton-Jacobi formalism in locally conformally symplectic manifolds. Our interest in the Hamilton-Jacobi theory comes from the suitability of this theory as an integration method for dynamical systems, whilst our interest in the locally conformal character will account for physical theories described by Hamiltonians defined on well-behaved line bundles, whose dynamic…
▽ More
In this article we provide a Hamilton-Jacobi formalism in locally conformally symplectic manifolds. Our interest in the Hamilton-Jacobi theory comes from the suitability of this theory as an integration method for dynamical systems, whilst our interest in the locally conformal character will account for physical theories described by Hamiltonians defined on well-behaved line bundles, whose dynamic takes place in open subsets of the general manifold. We present a local l.c.s. Hamilton-Jacobi in subsets of the general manifold, and then provide a global view by using the Lichnerowicz-deRham differential. We show a comparison between the global and local description of a l.c.s. Hamilton--Jacobi theory, and how actually the local behavior can be glued to retrieve the global behavior of the Hamilton-Jacobi theory.
△ Less
Submitted 4 October, 2019;
originally announced October 2019.
-
Google Research Football: A Novel Reinforcement Learning Environment
Authors:
Karol Kurach,
Anton Raichuk,
Piotr Stańczyk,
Michał Zając,
Olivier Bachem,
Lasse Espeholt,
Carlos Riquelme,
Damien Vincent,
Marcin Michalski,
Olivier Bousquet,
Sylvain Gelly
Abstract:
Recent progress in the field of reinforcement learning has been accelerated by virtual learning environments such as video games, where novel algorithms and ideas can be quickly tested in a safe and reproducible manner. We introduce the Google Research Football Environment, a new reinforcement learning environment where agents are trained to play football in an advanced, physics-based 3D simulator…
▽ More
Recent progress in the field of reinforcement learning has been accelerated by virtual learning environments such as video games, where novel algorithms and ideas can be quickly tested in a safe and reproducible manner. We introduce the Google Research Football Environment, a new reinforcement learning environment where agents are trained to play football in an advanced, physics-based 3D simulator. The resulting environment is challenging, easy to use and customize, and it is available under a permissive open-source license. In addition, it provides support for multiplayer and multi-agent experiments. We propose three full-game scenarios of varying difficulty with the Football Benchmarks and report baseline results for three commonly used reinforcement algorithms (IMPALA, PPO, and Ape-X DQN). We also provide a diverse set of simpler scenarios with the Football Academy and showcase several promising research directions.
△ Less
Submitted 14 April, 2020; v1 submitted 25 July, 2019;
originally announced July 2019.
-
Hamilton-Jacobi theory for gauge field theories
Authors:
Manuel de León,
Marcin Zając
Abstract:
Recently, M. de León el al. ([8]) have developed a geometrical description of Hamilton-Jacobi theory for multisymplectic field theory. In our paper we analyse in the same spirit a special kind of field theories which are gauge field theories. The Hamilton-Jacobi theory for this kind of fields is shown.
Recently, M. de León el al. ([8]) have developed a geometrical description of Hamilton-Jacobi theory for multisymplectic field theory. In our paper we analyse in the same spirit a special kind of field theories which are gauge field theories. The Hamilton-Jacobi theory for this kind of fields is shown.
△ Less
Submitted 23 April, 2019;
originally announced April 2019.
-
Split Batch Normalization: Improving Semi-Supervised Learning under Domain Shift
Authors:
Michał Zając,
Konrad Zolna,
Stanisław Jastrzębski
Abstract:
Recent work has shown that using unlabeled data in semi-supervised learning is not always beneficial and can even hurt generalization, especially when there is a class mismatch between the unlabeled and labeled examples. We investigate this phenomenon for image classification on the CIFAR-10 and the ImageNet datasets, and with many other forms of domain shifts applied (e.g. salt-and-pepper noise).…
▽ More
Recent work has shown that using unlabeled data in semi-supervised learning is not always beneficial and can even hurt generalization, especially when there is a class mismatch between the unlabeled and labeled examples. We investigate this phenomenon for image classification on the CIFAR-10 and the ImageNet datasets, and with many other forms of domain shifts applied (e.g. salt-and-pepper noise). Our main contribution is Split Batch Normalization (Split-BN), a technique to improve SSL when the additional unlabeled data comes from a shifted distribution. We achieve it by using separate batch normalization statistics for unlabeled examples. Due to its simplicity, we recommend it as a standard practice. Finally, we analyse how domain shift affects the SSL training process. In particular, we find that during training the statistics of hidden activations in late layers become markedly different between the unlabeled and the labeled examples.
△ Less
Submitted 6 April, 2019;
originally announced April 2019.
-
ZETH: On Integrating Zerocash on Ethereum
Authors:
Antoine Rondelet,
Michal Zajac
Abstract:
Transaction privacy is a hard problem on an account-based blockchain such as Ethereum. While Ben-Sasson et al. presented the Zerocash protocol [BCG+14] as a decentralized anonymous payment (DAP) scheme standing on top of Bitcoin, no study about the integration of such DAP on top of a ledger defined in the account model was provided. In this paper we aim to fill this gap and propose ZETH, an adapta…
▽ More
Transaction privacy is a hard problem on an account-based blockchain such as Ethereum. While Ben-Sasson et al. presented the Zerocash protocol [BCG+14] as a decentralized anonymous payment (DAP) scheme standing on top of Bitcoin, no study about the integration of such DAP on top of a ledger defined in the account model was provided. In this paper we aim to fill this gap and propose ZETH, an adaptation of Zerocash that can be deployed on top of Ethereum without making any change to the base layer. Our study shows that not only ZETH could be used to transfer Ether, the base currency of Ethereum, but it could also be used to transfer other types of smart contract-based digital assets. We propose an analysis of ZETH's privacy promises and argue that information leakages intrinsic to the use of this protocol are controlled and well-defined, which makes it a viable solution to support private transactions in the context of public and permissioned chains.
△ Less
Submitted 4 April, 2019; v1 submitted 1 April, 2019;
originally announced April 2019.
-
Site-selective quantum control in an isotopically enriched 28Si/SiGe quadruple quantum dot
Authors:
A. J. Sigillito,
J. C. Loy,
D. M. Zajac,
M. J. Gullans,
L. F. Edge,
J. R. Petta
Abstract:
Silicon spin qubits are a promising quantum computing platform offering long coherence times, small device sizes, and compatibility with industry-backed device fabrication techniques. In recent years, high fidelity single-qubit and two-qubit operations have been demonstrated in Si. Here, we demonstrate coherent spin control in a quadruple quantum dot fabricated using isotopically enriched 28Si. We…
▽ More
Silicon spin qubits are a promising quantum computing platform offering long coherence times, small device sizes, and compatibility with industry-backed device fabrication techniques. In recent years, high fidelity single-qubit and two-qubit operations have been demonstrated in Si. Here, we demonstrate coherent spin control in a quadruple quantum dot fabricated using isotopically enriched 28Si. We tune the ground state charge configuration of the quadruple dot down to the single electron regime and demonstrate tunable interdot tunnel couplings as large as 20 GHz, which enables exchange-based two-qubit gate operations. Site-selective single spin rotations are achieved using electric dipole spin resonance in a magnetic field gradient. We execute a resonant-CNOT gate between two adjacent spins in 270 ns.
△ Less
Submitted 14 March, 2019;
originally announced March 2019.
-
Adversarial Framing for Image and Video Classification
Authors:
Konrad Zolna,
Michal Zajac,
Negar Rostamzadeh,
Pedro O. Pinheiro
Abstract:
Neural networks are prone to adversarial attacks. In general, such attacks deteriorate the quality of the input by either slightly modifying most of its pixels, or by occluding it with a patch. In this paper, we propose a method that keeps the image unchanged and only adds an adversarial framing on the border of the image. We show empirically that our method is able to successfully attack state-of…
▽ More
Neural networks are prone to adversarial attacks. In general, such attacks deteriorate the quality of the input by either slightly modifying most of its pixels, or by occluding it with a patch. In this paper, we propose a method that keeps the image unchanged and only adds an adversarial framing on the border of the image. We show empirically that our method is able to successfully attack state-of-the-art methods on both image and video classification problems. Notably, the proposed method results in a universal attack which is very fast at test time. Source code can be found at https://github.com/zajaczajac/adv_framing .
△ Less
Submitted 17 October, 2019; v1 submitted 11 December, 2018;
originally announced December 2018.
-
Single-Spin Relaxation in a Synthetic Spin-Orbit Field
Authors:
F. Borjans,
D. M. Zajac,
T. M. Hazard,
J. R. Petta
Abstract:
Strong magnetic field gradients can produce a synthetic spin-orbit interaction that allows for high fidelity electrical control of single electron spins. We investigate how a field gradient impacts the spin relaxation time T_1 by measuring T_1 as a function of magnetic field B in silicon. The interplay of charge noise, magnetic field gradients, phonons, and conduction band valleys leads to a maxim…
▽ More
Strong magnetic field gradients can produce a synthetic spin-orbit interaction that allows for high fidelity electrical control of single electron spins. We investigate how a field gradient impacts the spin relaxation time T_1 by measuring T_1 as a function of magnetic field B in silicon. The interplay of charge noise, magnetic field gradients, phonons, and conduction band valleys leads to a maximum relaxation time of 160 ms at low field, a strong spin-valley relaxation hotspot at intermediate fields, and a B^4 scaling at high fields. T_1 is found to decrease with lattice temperature T_lat as well as with added electrical noise. In comparison, samples without micromagnets have a significantly longer T_1. Optimization of the micromagnet design, combined with reductions in charge noise and electron temperature, may further extend T_1 in devices with large magnetic field gradients.
△ Less
Submitted 2 November, 2018;
originally announced November 2018.