Search

The development of programmable quantum devices can be measured by the complexity of manybody states that they are able to prepare. Among the most significant are topologically ordered states of matter, which enable robust quantum information storage and processing. While topological orders are more readily accessible with qudits, experimental realisations have thus far been limited to lattice models of qubits. Here, we prepare a ground state of the Z3 toric code state on 24 qutrits in a trapped ion quantum processor with fidelity per qutrit exceeding 96.5(3)%. We manipulate two types of defects which go beyond the conventional qubit toric code: a parafermion, and its bound state which is related to charge conjugation symmetry. We further demonstrate defect fusion and the transfer of entanglement between anyons and defects, which we use to control topological qutrits. Our work opens up the space of long-range entangled states with qudit degrees of freedom for use in quantum simulation and universal error-correcting codes., Comment: 8+20 pages, 15 figures

Creating and manipulating anyons and symmetry defects in topological phases, especially those with a non-Abelian character, constitutes a primitive for topological quantum computation. We provide a physical protocol for implementing the ribbon operators of non-Abelian anyons and symmetry defects. We utilize dualities, in particular the Kramers-Wannier or gauging map, which have previously been used to construct topologically ordered ground states by relating them to simpler states. In this work, ribbon operators are implemented by applying a gauging procedure to a lower-dimensional region of such states. This protocol uses sequential unitary circuits or, in certain cases, constant-depth adaptive circuits. We showcase this for anyons and defects in the $\mathbb{Z}_3$ toric code and $S_3$ quantum double. The general applicability of our method is demonstrated by deriving unitary expressions for ribbon operators of various (twisted) quantum doubles., Comment: 7+18 pages, 3+8 figures

Large Language Models (LLMs) have propelled groundbreaking advancements across several domains and are commonly used for text generation applications. However, the computational demands of these complex models pose significant challenges, requiring efficient hardware acceleration. Benchmarking the performance of LLMs across diverse hardware platforms is crucial to understanding their scalability and throughput characteristics. We introduce LLM-Inference-Bench, a comprehensive benchmarking suite to evaluate the hardware inference performance of LLMs. We thoroughly analyze diverse hardware platforms, including GPUs from Nvidia and AMD and specialized AI accelerators, Intel Habana and SambaNova. Our evaluation includes several LLM inference frameworks and models from LLaMA, Mistral, and Qwen families with 7B and 70B parameters. Our benchmarking results reveal the strengths and limitations of various models, hardware platforms, and inference frameworks. We provide an interactive dashboard to help identify configurations for optimal performance for a given hardware platform.

Large language models have drastically changed the prospects of AI by introducing technologies for more complex natural language processing. However, current methodologies to train such LLMs require extensive resources including but not limited to large amounts of data, expensive machinery, and lengthy training. To solve this problem, this paper proposes a new tokenization method inspired by universal Lempel-Ziv-Welch data compression that compresses repetitive phrases into multi-word tokens. With MultiTok as a new tokenizing tool, we show that language models are able to be trained notably more efficiently while offering a similar accuracy on more succinct and compressed training data. In fact, our results demonstrate that MultiTok achieves a comparable performance to the BERT standard as a tokenizer while also providing close to 2.5x faster training with more than 30% less training data.

Stablecoins are a class of cryptocurrencies which aim at providing consistency and predictability, typically by pegging the token's value to that of a real world asset. Designing resilient decentralized stablecoins is a challenge, and prominent stablecoins today either (i) give up on decentralization, or (ii) rely on user-owned cryptocurrencies as collateral, exposing the token to exogenous price fluctuations. In this latter category, it is increasingly common to employ algorithmic mechanisms to automate risk management, helping maintain the peg. One example of this is Reflexer's RAI, which adapts its system-internal exchange rate (redemption price) to secondary market conditions according to a proportional control law. In this paper, we take this idea of active management a step further, and introduce a new kind of control scheme based on a Stackelberg game model between the token protocol and its users. By doing so, we show that (i) we can mitigate adverse depeg events that inevitably arise in a fixed-redemption scheme such as MakerDao's DAI and (ii) generally outperform a simpler, adaptive-redemption scheme such as RAI in the task of targeting a desired market price. We demonstrate these results through extensive simulations over a range of market conditions.

The identification of novel mechanisms for superconductivity is a longstanding goal in theoretical physics. In this work, we argue that the combination of repulsive interactions and high magnetic fields can generate electron pairing, phase coherence and superconductivity. Inspired by the large lattice constants of moir\'e materials, which make large flux per unit cell accessible at laboratory fields, we study the triangular lattice Hofstadter-Hubbard model at one-quarter flux quantum per plaquette, where previous literature has argued that a chiral spin liquid separates a weak-coupling integer quantum Hall phase and a strong-coupling topologically-trivial Mott insulator. We argue that topological superconductivity emerges upon doping in the vicinity of the integer quantum Hall to chiral spin liquid transition. We employ exact diagonalization and density matrix renormalization group methods to examine this theoretical scenario and find that electronic pairing indeed occurs above the half-filled ground states not just near the putative critical point but over a remarkably broad range of coupling strengths on both sides of criticality. On the chiral spin liquid side, our results provide a concrete model realization of the storied mechanism of anyon superconductivity. Our study thus establishes a beyond-BCS mechanism for electron pairing in a well-controlled limit, relying crucially on the interplay between electron correlations and band topology.

Language models (LMs) have demonstrated expert-level reasoning and recall abilities in medicine. However, computational costs and privacy concerns are mounting barriers to wide-scale implementation. We introduce a parsimonious adaptation of phi-3-mini, MedMobile, a 3.8 billion parameter LM capable of running on a mobile device, for medical applications. We demonstrate that MedMobile scores 75.7% on the MedQA (USMLE), surpassing the passing mark for physicians (~60%), and approaching the scores of models 100 times its size. We subsequently perform a careful set of ablations, and demonstrate that chain of thought, ensembling, and fine-tuning lead to the greatest performance gains, while unexpectedly retrieval augmented generation fails to demonstrate significant improvements, Comment: 13 pages, 5 figures (2 main, 3 supplementary)

This work develops a distributed graph neural network (GNN) methodology for mesh-based modeling applications using a consistent neural message passing layer. As the name implies, the focus is on enabling scalable operations that satisfy physical consistency via halo nodes at sub-graph boundaries. Here, consistency refers to the fact that a GNN trained and evaluated on one rank (one large graph) is arithmetically equivalent to evaluations on multiple ranks (a partitioned graph). This concept is demonstrated by interfacing GNNs with NekRS, a GPU-capable exascale CFD solver developed at Argonne National Laboratory. It is shown how the NekRS mesh partitioning can be linked to the distributed GNN training and inference routines, resulting in a scalable mesh-based data-driven modeling workflow. We study the impact of consistency on the scalability of mesh-based GNNs, demonstrating efficient scaling in consistent GNNs for up to O(1B) graph nodes on the Frontier exascale supercomputer.

In this study, we derive the asymptotic expressions for the electrostatic force between two charged spherical conductors in an electric field. Davis \cite{davis1964two} initially provided an expression for these forces, which are split into components along the X and Z directions. These forces are influenced by several factors, including the separation distance between the spheres, their sizes, their charges, the applied external electric field, and 10 specific force coefficients. These coefficients involve complex series expressions, and computing them requires summing up the 24 infinite series. To tackle the challenge of series convergence, previous researchers (\cite{lekner2012electrostatics}, \cite{lekner2013forces}, \cite{arp1977particle}) have used various methods to derive asymptotic expressions. In this work, the method of matched asymptotic expansions has been implemented to derive the asymptotic forms of these 24 series, specifically for the scenario involving two equal-sized conducting spheres in an external electric field. Accurate calculation of these series is crucial, especially when the distance between the spheres is very small, making the series calculations computationally intensive. We utilize the method proposed by Cox and Brenner \cite{Cox1967} to derive the asymptotic expressions and hence compute the 10 force coefficients and finally apply them to compute the electrostatic force for small separations. Additionally, we assess the accuracy of these asymptotic expressions by comparing them to the actual series and analyzing their deviations through percentage errors.

In today's world of advanced AI technologies, data management is a critical component of any AI/ML solution. Effective data management is vital for the creation and maintenance of high-quality, diverse datasets, which significantly enhance predictive capabilities and lead to smarter business solutions. In this work, we introduce MetaPix, a Data-centric AI platform offering comprehensive data management solutions specifically designed for unstructured data. MetaPix offers robust tools for data ingestion, processing, storage, versioning, governance, and discovery. The platform operates on four key concepts: DataSources, Datasets, Extensions and Extractors. A DataSource serves as MetaPix top level asset, representing a narrow-scoped source of data for a specific use. Datasets are MetaPix second level object, structured collections of data. Extractors are internal tools integrated into MetaPix's backend processing, facilitate data processing and enhancement. Additionally, MetaPix supports extensions, enabling integration with external third-party tools to enhance platform functionality. This paper delves into each MetaPix concept in detail, illustrating how they collectively contribute to the platform's objectives. By providing a comprehensive solution for managing and utilizing unstructured computer vision data, MetaPix equips organizations with a powerful toolset to develop AI applications effectively., Comment: Accepted @ The 22nd International Conference on Software Engineering Research & Practice

A graph neural network (GNN) approach is introduced in this work which enables mesh-based three-dimensional super-resolution of fluid flows. In this framework, the GNN is designed to operate not on the full mesh-based field at once, but on localized meshes of elements (or cells) directly. To facilitate mesh-based GNN representations in a manner similar to spectral (or finite) element discretizations, a baseline GNN layer (termed a message passing layer, which updates local node properties) is modified to account for synchronization of coincident graph nodes, rendering compatibility with commonly used element-based mesh connectivities. The architecture is multiscale in nature, and is comprised of a combination of coarse-scale and fine-scale message passing layer sequences (termed processors) separated by a graph unpooling layer. The coarse-scale processor embeds a query element (alongside a set number of neighboring coarse elements) into a single latent graph representation using coarse-scale synchronized message passing over the element neighborhood, and the fine-scale processor leverages additional message passing operations on this latent graph to correct for interpolation errors. Demonstration studies are performed using hexahedral mesh-based data from Taylor-Green Vortex flow simulations at Reynolds numbers of 1600 and 3200. Through analysis of both global and local errors, the results ultimately show how the GNN is able to produce accurate super-resolved fields compared to targets in both coarse-scale and multiscale model configurations.

The imposition of a global constraint of the conservation of total kinetic energy on a forced-dissipative Burgers equation yields a governing equation that is invariant under the time-reversal symmetry operation, $\{\mathcal{T}: t \to -t; u \to -u \}$, where $u$ is the velocity field. Moreover, the dissipation term gets strongly modified, as the viscosity is no longer a constant, but a fluctuating, state dependent quantity, which can even become negative in certain dynamical regimes. Despite these differences, the statistical properties of different dynamical regimes of the time-reversible Burgers equation and the standard forced-dssipative Burgers equation are equivalent, \`a la Gallavotti's conjecture of \textit{equivalence of nonequilibrium ensembles}. We show that the negative viscosity events occur only in the thermalized regime described by the time-reversible equation. This quasi-equilibrium regime is examined by calculating the local Lyapunov spectra and fluctuation relations. A pairing symmetry among the spectra is observed, indicating that the dynamics is chaotic and has an attractor spanning the entire phase space of the system. The violations of the second law of thermodynamics are found to be in accordance with the fluctuation relations, namely the Gallavotti-Cohen relation based on the phase-space contraction rate and the Cohen-Searles fluctuation relation based on the energy production rate. It is also argued that these violations are associated with the effects of the Galerkin-truncation, the latter is responsible for the thermalization.

Twisted bilayer graphene (TBG) has elements in common with two paradigmatic examples of strongly correlated physics: quantum Hall physics and Hubbard physics. On one hand, TBG hosts flat topological Landau-level-like bands which exhibits quantum anomalous Hall effects. On the other hand, these bands have concentrated charge density and show signs of extensive entropy resembling local moments. The combination of these features leads to a question: can decoupled moments emerge in an isolated topological band, despite the lack of exponentially localized Wannier states? In this work, we answer the question affirmatively by proposing a minimal model for these bands in TBG that combines topology and charge concentration at the AA sites, leading to analytic wavefunctions that closely approximate those of the BM model with realistic parameters. Importantly, charge concentration also leads to Berry curvature concentration at $\Gamma$, generating a small parameter $s$ that yields analytic tractability. We show that, rather surprisingly, the model hosts nearly decoupled flavor moments without any extra degrees of freedom. These moments are non-local due to topology-enforced power-law tails, yet have parametrically small overlap. We further develop a diagrammatic expansion in which the self energy can be computed exactly to leading order in $s^2$ in the fluctuating moment regime. At charge neutrality, we find a "Mott semimetal", with large flavor entropy and a Mott gap everywhere in the BZ except for the vicinity of the $\Gamma$ point. Away from neutrality, the Mott semimetal gaps out in a spectrally imbalanced manner, with one Mott band having zero $Z_k$ at the $\Gamma$ point. The model accurately reproduces results from finite temperature thermodynamic measurements, leads to new experimental predictions, and resolves the problem of the emergence of Hubbard physics in isolated topological bands., Comment: 22 pages

In this paper, we consider a K-user interference channel where interference among the users is neither too strong nor too weak, a scenario that is relatively underexplored in the literature. We propose a novel deep learning-based approach to design the encoder and decoder functions that aim to maximize the sumrate of the interference channel for discrete constellations. We first consider the MaxSINR algorithm, a state-of-the-art linear scheme for Gaussian inputs, as the baseline and then propose a modified version of the algorithm for discrete inputs. We then propose a neural network-based approach that learns a constellation mapping with the objective of maximizing the sumrate. We provide numerical results to show that the constellations learned by the neural network-based approach provide enhanced alignments, not just in beamforming directions but also in terms of the effective constellation at the receiver, thereby leading to improved sum-rate performance.

Machine ethics is the field that studies how ethical behaviour can be accomplished by autonomous systems. While there exist some systematic reviews aiming to consolidate the state of the art in machine ethics prior to 2020, these tend to not include work that uses reinforcement learning agents as entities whose ethical behaviour is to be achieved. The reason for this is that only in the last years we have witnessed an increase in machine ethics studies within reinforcement learning. We present here a systematic review of reinforcement learning for machine ethics and machine ethics within reinforcement learning. Additionally, we highlight trends in terms of ethics specifications, components and frameworks of reinforcement learning, and environments used to result in ethical behaviour. Our systematic review aims to consolidate the work in machine ethics and reinforcement learning thus completing the gap in the state of the art machine ethics landscape

At the core of both successful generative and self-supervised representation learning models there is a reconstruction objective that incorporates some form of image corruption. Diffusion models implement this approach through a scheduled Gaussian corruption process, while masked auto-encoder models do so by masking patches of the image. Despite their different approaches, the underlying similarity in their methodologies suggests a promising avenue for an auto-encoder capable of both de-noising tasks. We propose a unified self-supervised objective, dubbed Unified Masked Diffusion (UMD), that combines patch-based and noise-based corruption techniques within a single auto-encoding framework. Specifically, UMD modifies the diffusion transformer (DiT) training process by introducing an additional noise-free, high masking representation step in the diffusion noising schedule, and utilizes a mixed masked and noised image for subsequent timesteps. By integrating features useful for diffusion modeling and for predicting masked patch tokens, UMD achieves strong performance in downstream generative and representation learning tasks, including linear probing and class-conditional generation. This is achieved without the need for heavy data augmentations, multiple views, or additional encoders. Furthermore, UMD improves over the computational efficiency of prior diffusion based methods in total training time. We release our code at https://github.com/philippe-eecs/small-vision., Comment: 19 Pages, 8 Figures, 3Tables

We introduce OpenDebateEvidence, a comprehensive dataset for argument mining and summarization sourced from the American Competitive Debate community. This dataset includes over 3.5 million documents with rich metadata, making it one of the most extensive collections of debate evidence. OpenDebateEvidence captures the complexity of arguments in high school and college debates, providing valuable resources for training and evaluation. Our extensive experiments demonstrate the efficacy of fine-tuning state-of-the-art large language models for argumentative abstractive summarization across various methods, models, and datasets. By providing this comprehensive resource, we aim to advance computational argumentation and support practical applications for debaters, educators, and researchers. OpenDebateEvidence is publicly available to support further research and innovation in computational argumentation. Access it here: https://huggingface.co/datasets/Yusuf5/OpenCaselist, Comment: Published to the 38th Conference on Neural Information Processing Systems (NeurIPS 2024) Track on Datasets and Benchmarks

The robustness of the macroscopic quantum nature of a superconductor can be characterized by the superfluid stiffness, $\rho_s$, a quantity that describes the energy required to vary the phase of the macroscopic quantum wave function. In unconventional superconductors, such as cuprates, the low-temperature behavior of $\rho_s$ drastically differs from that of conventional superconductors due to quasiparticle excitations from gapless points (nodes) in momentum space. Intensive research on the recently discovered magic-angle twisted graphene family has revealed, in addition to superconducting states, strongly correlated electronic states associated with spontaneously broken symmetries, inviting the study of $\rho_s$ to uncover the potentially unconventional nature of its superconductivity. Here we report the measurement of $\rho_s$ in magic-angle twisted trilayer graphene (TTG), revealing unconventional nodal-gap superconductivity. Utilizing radio-frequency reflectometry techniques to measure the kinetic inductive response of superconducting TTG coupled to a microwave resonator, we find a linear temperature dependence of $\rho_s$ at low temperatures and nonlinear Meissner effects in the current bias dependence, both indicating nodal structures in the superconducting order parameter. Furthermore, the doping dependence shows a linear correlation between the zero temperature $\rho_s$ and the superconducting transition temperature $T_c$, reminiscent of Uemura's relation in cuprates, suggesting phase-coherence-limited superconductivity. Our results provide strong evidence for nodal superconductivity in TTG and put strong constraints on the mechanisms of these graphene-based superconductors.

For a given parent fractional quantum Hall (FQH) state at filling fraction $\nu$, the hierarchy construction produces FQH states at nearby filling fractions $\{\nu_n\}$ by condensing minimally charged quasiholes or quasiparticles of the parent state into their own FQH states. The hierarchy construction has been useful for relating families of FQH states and for the experimental identification of the topological order of parent states via the presence of daughter states. We reinterpret the hierarchy construction as a two-step procedure: stacking with a second FQH state and condensing a condensable algebra of bosons. This two-step procedure can be applied to both abelian and non-abelian FQH states, and it does not require calculations with a wavefunction. We show this construction reproduces the hierarchies for the Laughlin and Pfaffian states, and can be applied further to propose hierarchies for various non-abelian FQH states., Comment: 28 pages, 1 figure

While automatic speech recognition (ASR) greatly benefits from data augmentation, the augmentation recipes themselves tend to be heuristic. In this paper, we address one of the heuristic approach associated with balancing the right amount of augmented data in ASR training by introducing a reinforcement learning (RL) based dynamic adjustment of original-to-augmented data ratio (OAR). Unlike the fixed OAR approach in conventional data augmentation, our proposed method employs a deep Q-network (DQN) as the RL mechanism to learn the optimal dynamics of OAR throughout the wav2vec2.0 based ASR training. We conduct experiments using the LibriSpeech dataset with varying amounts of training data, specifically, the 10Min, 1H, 10H, and 100H splits to evaluate the efficacy of the proposed method under different data conditions. Our proposed method, on average, achieves a relative improvement of 4.96% over the open-source wav2vec2.0 base model on standard LibriSpeech test sets., Comment: Accepted: Interspeech 2024

Square lattice Hubbard models with tunable hopping ratio $t'/t$ are highly promising for realizing a variety of quantum phases and for shedding light on key puzzles in correlated quantum materials, including higher-temperature superconductivity. We show that twisted square lattice homo-bilayers generically offer such tunability when the flat bands originate from the corner of the Brillouin zone. We reveal an emergent symmetry at low twist-angles, absent in single layers, that necessitates the vanishing of nearest neighbor hopping ($t=0$). This symmetry can be lifted by an inter-layer displacement field or by an in-plane magnetic field, introducing tunable $t$ and anisotropy, allowing access to a wide range of $t'/t$ ratios for correlated electrons on a moir\'e square lattice., Comment: 5 pages + Supplementary

Motivated by recent experiments realizing long-lived non-equilibrium states in aperiodically driven quantum many-body systems, we investigate the dynamics of a quasiperiodically driven Rydberg atom chain in the strong Rydberg blockage regime. In this regime, the system is kinetically constrained and the `PXP' model describes its dynamics. Even without driving, the PXP model exhibits many-body scarring and resultant persistent oscillations for dynamics originating from the N\'{e}el-ordered initial state. We demonstrate that a rich array of dynamical behaviors emerge when the system is subjected to a continuous drive. In the high-frequency regime, the system exhibits revivals and oscillations for the N\'{e}el ordered initial state both for periodic and quasi-periodic drives. We trace the origin of this non-ergodicity to an effective PXP Hamiltonian for both of these driving protocols in this regime. Furthermore, we demonstrate that the behavior of the fidelity and the entanglement entropy is non-monotonic at low frequencies in the high-amplitude regime. This leads to several re-entrant scarring transitions both for both the N\'{e}el-ordered and the fully polarized initial state. Our results demonstrate that continuous quasi-periodic drive protocols can provide a promising route to realize prethermal phases of matter in kinetically constrained systems., Comment: 14 pages, 6 figures

We propose a systematic wave function based approach to construct topological invariants for families of lattice systems that are short-range entangled using local parameter spaces. This construction is particularly suitable when given a family of tensor networks that can be viewed as the ground states of $d$ dimensional lattice systems, for which we construct the closed $(d+2)$-form higher Berry curvature, which is a generalization of the well known 2-form Berry curvature. Such $(d+2)$-form higher Berry curvature characterizes a flow of $(d+1)$-form higher Berry curvature in the system. Our construction is equally suitable for constructing other higher pumps, such as the (higher) Thouless pump in the presence of a global on-site $U(1)$ symmetry, which corresponds to a closed $d$-form. The cohomology classes of such higher differential forms are topological invariants and are expected to be quantized for short-range entangled states. We illustrate our construction with exactly solvable lattice models that are in nontrivial higher Berry classes in $d=2$., Comment: 18 pages, 4 figures

Higher Berry curvature (HBC) is the proposed generalization of Berry curvature to infinitely extended systems. Heuristically HBC captures the flow of local Berry curvature in a system. Here we provide a simple formula for computing the HBC for extended $d = 1$ systems at the level of wave functions using the Schmidt decomposition. We also find a corresponding formula for matrix product states (MPS), and show that for translationally invariant MPS this gives rise to a quantized invariant. We demonstrate our approach with an exactly solvable model and numerical calculations for generic models using iDMRG, Comment: 9 pages, 2 figures

In multilayer moir\'e heterostructures, the interference of multiple twist angles ubiquitously leads to tunable ultra-long-wavelength patterns known as supermoir\'e lattices. However, their impact on the system's many-body electronic phase diagram remains largely unexplored. We present local compressibility measurements revealing numerous incompressible states resulting from supermoir\'e-lattice-scale isospin symmetry breaking driven by strong interactions. By using the supermoir\'e lattice occupancy as a probe of isospin symmetry, we observe an unexpected doubling of the miniband filling near $\nu=-2$, possibly indicating a hidden phase transition or normal-state pairing proximal to the superconducting phase. Our work establishes supermoir\'e lattices as a tunable parameter for designing novel quantum phases and an effective tool for unraveling correlated phenomena in moir\'e materials.

The paper presents a novel Auto Language Prediction Dictionary Capsule (ALPDC) framework for language prediction and machine translation. The model uses a combination of neural networks and symbolic representations to predict the language of a given input text and then translate it to a target language using pre-built dictionaries. This research work also aims to translate the text of various languages to its literal meaning in English. The proposed model achieves state-of-the-art results on several benchmark datasets and significantly improves translation accuracy compared to existing methods. The results show the potential of the proposed method for practical use in multilingual communication and natural language processing tasks., Comment: 21 Pages

Conversational systems are crucial for human-computer interaction, managing complex dialogues by identifying threads and prioritising responses. This is especially vital in multi-party conversations, where precise identification of threads and strategic response prioritisation ensure efficient dialogue management. To address these challenges an end-to-end model that identifies threads and prioritises their response generation based on the importance was developed, involving a systematic decomposition of the problem into discrete components - thread detection, prioritisation, and performance optimisation which was meticulously analysed and optimised. These refined components seamlessly integrate into a unified framework, in conversational systems. Llama2 7b is used due to its high level of generalisation but the system can be updated with any open source Large Language Model(LLM). The computational capabilities of the Llama2 model was augmented by using fine tuning methods and strategic prompting techniques to optimise the model's performance, reducing computational time and increasing the accuracy of the model. The model achieves up to 10x speed improvement, while generating more coherent results compared to existing models., Comment: 6 pages, 4 figures, submitted to 2024 IEEE International Conference on Signal Processing and Communications (SPCOM)

We propose a minimal "three-patch model" for the anomalous Hall crystal (AHC), a topological electronic state that spontaneously breaks both time-reversal symmetry and continuous translation symmetry. The proposal for this state is inspired by the recently observed integer and fractional quantum Hall states in rhombohedral multilayer graphene at zero magnetic field. There, interaction effects appear to amplify the effects of a weak moir\'e potential, leading to the formation of stable, isolated Chern bands. It has been further shown that Chern bands are stabilized in mean field calculations even without a moir\'e potential, enabling a realization of the AHC state. Our model is built upon the dissection of the Brillouin zone into patches centered around high symmetry points. Within this model, the wavefunctions at high symmetry points fully determine the topology and energetics of the state. We extract two quantum geometrical phases of the non-interacting wavefunctions that control the stability of the topologically nontrivial AHC state. The model predicts that the AHC state wins over the topological trivial Wigner crystal in a wide range of parameters, and agrees very well with the results of full self-consistent Hartree-Fock calculations of the rhombohedral multilayer graphene Hamiltonian., Comment: 15+10 pages, 9+5 figures

We propose a variant of Hamiltonian Monte Carlo (HMC), called the Repelling-Attracting Hamiltonian Monte Carlo (RAHMC), for sampling from multimodal distributions. The key idea that underpins RAHMC is a departure from the conservative dynamics of Hamiltonian systems, which form the basis of traditional HMC, and turning instead to the dissipative dynamics of conformal Hamiltonian systems. In particular, RAHMC involves two stages: a mode-repelling stage to encourage the sampler to move away from regions of high probability density; and, a mode-attracting stage, which facilitates the sampler to find and settle near alternative modes. We achieve this by introducing just one additional tuning parameter -- the coefficient of friction. The proposed method adapts to the geometry of the target distribution, e.g., modes and density ridges, and can generate proposals that cross low-probability barriers with little to no computational overhead in comparison to traditional HMC. Notably, RAHMC requires no additional information about the target distribution or memory of previously visited modes. We establish the theoretical basis for RAHMC, and we discuss repelling-attracting extensions to several variants of HMC in literature. Finally, we provide a tuning-free implementation via dual-averaging, and we demonstrate its effectiveness in sampling from, both, multimodal and unimodal distributions in high dimensions., Comment: 41 pages, 10 figures, 4 tables

The rise of moir\'{e} materials has led to experimental realizations of integer and fractional Chern insulators in small or vanishing magnetic fields. At the same time, a set of minimal conditions sufficient to guarantee a Abelian fractional state in a flat band were identified, namely "ideal" or "vortexable" quantum geometry. Such vortexable bands share essential features with the lowest Landau level, while excluding the need for more fine-tuned aspects such as flat Berry curvature. A natural and important generalization is to ask if such conditions can be extended to capture the quantum geometry of higher Landau levels, particularly the first (1LL), where non-Abelian states at $\nu = 1/2,2/5$ are known to be competitive. The possibility of realizing these states at zero magnetic field , and perhaps even more exotic ones, could become a reality if we could identify the essential structure of the 1LL in Chern bands. In this work, we introduce a precise definition of 1LL quantum geometry, along with a figure of merit that measures how closely a given band approaches the 1LL. We apply the definition to identify two models with 1LL structure -- a toy model of double bilayer twisted graphene and a more realistic model of strained Bernal graphene., Comment: 12 pages, 5 figures

The accuracy of modern automatic speaker verification (ASV) systems, when trained exclusively on adult data, drops substantially when applied to children's speech. The scarcity of children's speech corpora hinders fine-tuning ASV systems for children's speech. Hence, there is a timely need to explore more effective ways of reusing adults' speech data. One promising approach is to align vocal-tract parameters between adults and children through children-specific data augmentation, referred here to as ChildAugment. Specifically, we modify the formant frequencies and formant bandwidths of adult speech to emulate children's speech. The modified spectra are used to train ECAPA-TDNN (emphasized channel attention, propagation, and aggregation in time-delay neural network) recognizer for children. We compare ChildAugment against various state-of-the-art data augmentation techniques for children's ASV. We also extensively compare different scoring methods, including cosine scoring, PLDA (probabilistic linear discriminant analysis), and NPLDA (neural PLDA). We also propose a low-complexity weighted cosine score for extremely low-resource children ASV. Our findings on the CSLU kids corpus indicate that ChildAugment holds promise as a simple, acoustics-motivated approach, for improving state-of-the-art deep learning based ASV for children. We achieve up to 12.45% (boys) and 11.96% (girls) relative improvement over the baseline., Comment: The following article has been accepted by The Journal of the Acoustical Society of America (JASA). After it is published, it will be found at https://pubs.aip.org/asa/jasa

With the exponential growth in data volume and the emergence of data-intensive applications, particularly in the field of machine learning, concerns related to resource utilization, privacy, and fairness have become paramount. This paper focuses on the textual domain of data and addresses challenges regarding encoding sentences to their optimized representations through the lens of information-theory. In particular, we use empirical estimates of mutual information, using the Donsker-Varadhan definition of Kullback-Leibler divergence. Our approach leverages this estimation to train an information-theoretic sentence embedding, called TexShape, for (task-based) data compression or for filtering out sensitive information, enhancing privacy and fairness. In this study, we employ a benchmark language model for initial text representation, complemented by neural networks for information-theoretic compression and mutual information estimations. Our experiments demonstrate significant advancements in preserving maximal targeted information and minimal sensitive information over adverse compression ratios, in terms of predictive accuracy of downstream models that are trained using the compressed data., Comment: Submitted to the 2024 IEEE International Symposium on Information Theory

Cloud gaming, wherein game graphics is rendered in the cloud and streamed back to the user as real-time video, expands the gaming market to billions of users who do not have gaming consoles or high-power graphics PCs. Companies like Nvidia, Amazon, Sony and Microsoft are investing in building cloud gaming platforms to tap this large unserved market. However, cloud gaming requires the user to have high bandwidth and stable network connectivity - whereas a typical console game needs about 100-200 kbps, a cloud game demands minimum 10-20 Mbps. This makes the Internet Service Provider (ISP) a key player in ensuring the end-user's good gaming experience. In this paper we develop a method to detect Nvidia's GeForce NOW cloud gaming sessions over their network infrastructure, and measure associated user experience. In particular, we envision ISPs taking advantage of our method to provision network capacity at the right time and in the right place to support growth in cloud gaming at the right experience level; as well as identify the role of contextual factors such as user setup (browser vs app) and connectivity type (wired vs wireless) in performance degradation. We first present a detailed anatomy of flow establishment and volumetric profiles of cloud gaming sessions over multiple platforms, followed by a method to detect gameplay and measure key experience aspects such as latency, frame rate and resolution via real-time analysis of network traffic. The insights and methods are also validated in the lab for XBox Cloud Gaming platform. We then implement and deploy our method in a campus network to capture gameplay behaviors and experience measures across various user setups and connectivity types which we believe are valuable for network operators., Comment: This paper is accepted at Passive and Active Measurement (PAM) conference Mar 2024