Search

It is now a common business practice to buy access to large language model (LLM) inference rather than self-host, because of significant upfront hardware infrastructure and energy costs. However, as a buyer, there is no mechanism to verify the authenticity of the advertised service including the serving hardware platform, e.g. that it is actually being served using an NVIDIA H100. Furthermore, there are reports suggesting that model providers may deliver models that differ slightly from the advertised ones, often to make them run on less expensive hardware. That way, a client pays premium for a capable model access on more expensive hardware, yet ends up being served by a (potentially less capable) cheaper model on cheaper hardware. In this paper we introduce \textit{\textbf{hardware and software platform inference (HSPI)}} -- a method for identifying the underlying \GPU{} architecture and software stack of a (black-box) machine learning model solely based on its input-output behavior. Our method leverages the inherent differences of various \GPU{} architectures and compilers to distinguish between different \GPU{} types and software stacks. By analyzing the numerical patterns in the model's outputs, we propose a classification framework capable of accurately identifying the \GPU{} used for model inference as well as the underlying software configuration. Our findings demonstrate the feasibility of inferring \GPU{} type from black-box models. We evaluate HSPI against models served on different real hardware and find that in a white-box setting we can distinguish between different \GPU{}s with between $83.9\%$ and $100\%$ accuracy. Even in a black-box setting we are able to achieve results that are up to three times higher than random guess accuracy.

Based on the hyperboloidal framework, we research the dynamical process of charged de Sitter black holes scattered by a charged scalar field. From the linear perturbation analysis, with the coupling strength within a critical interval, the charged scalar field with a superradiance frequency can induce the instability of the system. To reveal the real-time dynamics of such an instability, the nonlinear numerical simulation is implemented. The results show that the scalar field grows exponentially in the early stages and drastically extracts the charge from the black hole due to the superradiance, analogous to the charged black hole in a closed system. Differently, after saturation, the scalar field can not coexist with the central black hole stably and dissipates beyond the cosmological horizon slowly, leaving behind a bald black hole., Comment: 16 pages, 12 figures

Implementing and tuning chirality is fundamental in physics, chemistry, and material science. Chiral charge density waves (CDWs), where chirality arises from correlated charge orders, are attracting intense interest due to their exotic transport and optical properties. However, a general framework for predicting chiral CDW materials is lacking, primarily because the underlying mechanisms remain elusive. Here, we address this challenge by developing the first comprehensive predictive framework, systematically identifying chiral CDW materials via first-principles calculations. The key lies in the previously overlooked phase difference of the CDW Q-vectors between layers, which is linked to opposite collective atomic displacements across different layers. This phase difference induces a spiral arrangement of the Q-vectors, ultimately giving rise to a chiral structure in real space. We validate our framework by applying it to the kagome lattice AV$_{3}$Sb$_{5}$ (A = K, Rb, Cs), successfully predicting emergent structural chirality. To demonstrate the generality of our approach, we extend it to predict chiral CDWs in the triangular-lattice NbSe$_{2}$. Beyond material predictions, our theory uncovers a universal and unprecedented Hall effect in chiral CDW materials, occurring without external magnetic fields or intrinsic magnetization. Our experiments on CsV$_{3}$Sb$_{5}$ confirm this prediction, observing a unique signature where the Hall conductivity's sign reverses when the input current is reversed, a phenomenon distinct from known Hall effects. Our findings elucidate the mechanisms behind chiral CDWs and open new avenues for discovering materials with unconventional quantum properties, with potential applications in next-generation electronic and spintronic devices.

In this paper, a two-stage intelligent scheduler is proposed to minimize the packet-level delay jitter while guaranteeing delay bound. Firstly, Lyapunov technology is employed to transform the delay-violation constraint into a sequential slot-level queue stability problem. Secondly, a hierarchical scheme is proposed to solve the resource allocation between multiple base stations and users, where the multi-agent reinforcement learning (MARL) gives the user priority and the number of scheduled packets, while the underlying scheduler allocates the resource. Our proposed scheme achieves lower delay jitter and delay violation rate than the Round-Robin Earliest Deadline First algorithm and MARL with delay violation penalty.

At least 25 kinds of detector-like devices need to be integrated in Phase I of the High Energy Photon Source (HEPS), and the work needs to be carefully planned to maximise productivity with highly limited human resources. After a systematic analysis on the actual work involved in detector integration, a separation of concerns between collaborating groups of personnel is established to minimise the duplication of efforts. To facilitate software development for detector integration, the ADGenICam library, which abstracts repeated code in EPICS modules for cameras, is extended to support a much wider range of detectors. An increasingly considerable fraction of detectors, both inside and outside HEPS, offer performance that exceed capabilities of the areaDetector framework in EPICS. Given this background, areaDetector's limitations in performance and architecture are analysed, and a QueueIOC -based framework that overcomes these limitations is introduced. A simple, flexible ZeroMQ-based protocol is used for data transport in this framework, while RDMA transport and multi-node readout will be explored for higher data throughputs. By calling C/C++ libraries from within Python, the performance of the former and the expressiveness of the latter can coexist nicely; the expressiveness allows for much higher efficiency in the implementation and use of integration modules functionally comparable to their EPICS counterparts., Comment: 11 pages, 3 figures

This study addresses the boundary artifacts in machine-learned (ML) parameterizations for ocean subgrid mesoscale momentum forcing, as identified in the online ML implementation from a previous study (Zhang et al., 2023). We focus on the boundary condition (BC) treatment within the existing convolutional neural network (CNN) models and aim to mitigate the "out-of-sample" errors observed near complex coastal regions without developing new, complex network architectures. Our approach leverages two established strategies for placing BCs in CNN models, namely zero and replicate padding. Offline evaluations revealed that these padding strategies significantly reduce root mean squared error (RMSE) in coastal regions by limiting the dependence on random initialization of weights and restricting the range of out-of-sample predictions. Further online evaluations suggest that replicate padding consistently reduces boundary artifacts across various retrained CNN models. In contrast, zero padding sometimes intensifies artifacts in certain retrained models despite both strategies performing similarly in offline evaluations. This study underscores the need for BC treatments in CNN models trained on open water data when predicting near-coastal subgrid forces in ML parameterizations. The application of replicate padding, in particular, offers a robust strategy to minimize the propagation of extreme values that can contaminate computational models or cause simulations to fail. Our findings provide insights for enhancing the accuracy and stability of ML parameterizations in the online implementation of ocean circulation models with coastlines.

Recently, through a unified gradient flow perspective of Markov chain Monte Carlo (MCMC) and variational inference (VI), particle-based variational inference methods (ParVIs) have been proposed that tend to combine the best of both worlds. While typical ParVIs such as Stein Variational Gradient Descent (SVGD) approximate the gradient flow within a reproducing kernel Hilbert space (RKHS), many attempts have been made recently to replace RKHS with more expressive function spaces, such as neural networks. While successful, these methods are mainly designed for sampling from unconstrained domains. In this paper, we offer a general solution to constrained sampling by introducing a boundary condition for the gradient flow which would confine the particles within the specific domain. This allows us to propose a new functional gradient ParVI method for constrained sampling, called constrained functional gradient flow (CFG), with provable continuous-time convergence in total variation (TV). We also present novel numerical strategies to handle the boundary integral term arising from the domain constraints. Our theory and experiments demonstrate the effectiveness of the proposed framework., Comment: NeurIPS 2024 camera-ready (30 pages, 26 figures)

Particle-based variational inference methods (ParVIs) use non-parametric variational families represented by particles to approximate the target distribution according to the kernelized Wasserstein gradient flow for the Kullback-Leibler (KL) divergence. Recent works introduce functional gradient flows to substitute the kernel for better flexibility. However, the deterministic updating mechanism may suffer from limited exploration and require expensive repetitive runs for new samples. In this paper, we propose Semi-Implicit Functional Gradient flow (SIFG), a functional gradient ParVI method that uses perturbed particles as the approximation family. The corresponding functional gradient flow, which can be estimated via denoising score matching, exhibits strong theoretical convergence guarantee. We also present an adaptive version of our method to automatically choose the suitable noise magnitude. Extensive experiments demonstrate the effectiveness and efficiency of the proposed framework on both simulated and real data problems., Comment: 31 pages, 12 figures

Metallic materials under high stress often exhibit deformation localization, manifesting as slip banding. Over seven decades ago, Frank and Read introduced the well-known model of dislocation multiplication at a source, explaining slip band formation. Here, we reveal two distinct types of slip bands (confined and extended) in alloys through multi-scale testing and modeling from microscopic to atomic scales. The confined slip band, characterized by a thin glide zone, arises from the conventional process of repetitive full dislocation emissions at Frank-Read source. Contrary to the classical model, the extended band stems from slip-induced deactivation of dislocation sources, followed by consequent generation of new sources on adjacent planes, leading to rapid band thickening. Our findings provide critical insights into atomic-scale collective dislocation motion and microscopic deformation instability in advanced structural materials, marking a pivotal advancement in our fundamental understanding of deformation dynamics., Comment: 17 pafes, 5 figures

We present a comprehensive study of spontaneous vectorization in the Einstein-Maxwell-Vector (EMV) model, uncovering a novel vectorization mechanism driven by the competition between electromagnetic and vector fields. By utilizing a generalized coordinate transformation, we resolve previous divergences near the event horizon and significantly expand the domain of existence for vectorized Reissner-Nordstr\"om black holes (VRNBHs). Our results highlight the crucial role of a newly defined combined charge $ \sqrt{\tilde{Q}^2 + \tilde{P}^2} $, which encapsulates this competition and establishes a clear thermodynamic preference, and a more natural description of light rights for VRNBHs over standard Reissner-Nordstr\"om black holes. Moreover, we identify unique features in the light ring structure of VRNBHs, revealing deeper connections to multi-charge black holes and opening new avenues for observational signatures in effective field theory., Comment: 31 pages, 11 figures

Recent success of diffusion models has inspired a surge of interest in developing sampling techniques using reverse diffusion processes. However, accurately estimating the drift term in the reverse stochastic differential equation (SDE) solely from the unnormalized target density poses significant challenges, hindering existing methods from achieving state-of-the-art performance. In this paper, we introduce the Diffusion-PINN Sampler (DPS), a novel diffusion-based sampling algorithm that estimates the drift term by solving the governing partial differential equation of the log-density of the underlying SDE marginals via physics-informed neural networks (PINN). We prove that the error of log-density approximation can be controlled by the PINN residual loss, enabling us to establish convergence guarantees of DPS. Experiments on a variety of sampling tasks demonstrate the effectiveness of our approach, particularly in accurately identifying mixing proportions when the target contains isolated components., Comment: 33 pages, 7 figures

In this paper, we investigate the problem of jamming detection and channel estimation during multi-user uplink beam training under random pilot jamming attacks in beamspace massive multi-input-multi-output (MIMO) systems. For jamming detection, we distinguish the signals from the jammer and the user by projecting the observation signals onto the pilot space. By using the multiple projected observation vectors corresponding to the unused pilots, we propose a jamming detection scheme based on the locally most powerful test (LMPT) for systems with general channel conditions. Analytical expressions for the probability of detection and false alarms are derived using the second-order statistics and likelihood functions of the projected observation vectors. For the detected jammer along with users, we propose a two-step minimum mean square error (MMSE) channel estimation using the projected observation vectors. As a part of the channel estimation, we develop schemes to estimate the norm and the phase of the inner-product of the legitimate pilot vector and the random jamming pilot vector, which can be obtained using linear MMSE estimation and a bilinear form of the multiple projected observation vectors. From simulations under different system parameters, we observe that the proposed technique improves the detection probability by 32.22% compared to the baseline at medium channel correlation level, and the channel estimation achieves a mean square error of -15.93dB., Comment: 13 pages, 9 figures. The paper has been submitted to an IEEE journal for possible publication

Post-training quantization of Large Language Models (LLMs) has proven effective in reducing the computational requirements for running inference on these models. In this study, we focus on a straightforward question: When aiming for a specific accuracy or perplexity target for low-precision quantization, how many high-precision numbers or calculations are required to preserve as we scale LLMs to larger sizes? We first introduce a critical metric named the quantization ratio, which compares the number of parameters quantized to low-precision arithmetic against the total parameter count. Through extensive and carefully controlled experiments across different model families, arithmetic types, and quantization granularities (e.g. layer-wise, matmul-wise), we identify two central phenomenons. 1) The larger the models, the better they can preserve performance with an increased quantization ratio, as measured by perplexity in pre-training tasks or accuracy in downstream tasks. 2) The finer the granularity of mixed-precision quantization (e.g., matmul-wise), the more the model can increase the quantization ratio. We believe these observed phenomena offer valuable insights for future AI hardware design and the development of advanced Efficient AI algorithms.

With the increasing acquisition of datasets over time, we now have access to precise and varied descriptions of the world, capturing all sorts of phenomena. These datasets can be seen as empirical observations of unknown causal generative processes, which can commonly be described by Structural Causal Models (SCMs). Recovering these causal generative processes from observations poses formidable challenges, and often require to learn a specific generative model for each dataset. In this work, we propose to learn a \emph{single} model capable of inferring in a zero-shot manner the causal generative processes of datasets. Rather than learning a specific SCM for each dataset, we enable the Fixed-Point Approach (FiP) proposed in~\cite{scetbon2024fip}, to infer the generative SCMs conditionally on their empirical representations. More specifically, we propose to amortize the learning of a conditional version of FiP to infer generative SCMs from observations and causal structures on synthetically generated datasets. We show that our model is capable of predicting in zero-shot the true generative SCMs, and as a by-product, of (i) generating new dataset samples, and (ii) inferring intervened ones. Our experiments demonstrate that our amortized procedure achieves performances on par with SoTA methods trained specifically for each dataset on both in and out-of-distribution problems. To the best of our knowledge, this is the first time that SCMs are inferred in a zero-shot manner from observations, paving the way for a paradigmatic shift towards the assimilation of causal knowledge across datasets.

he growing number of parameters and computational demands of large language models (LLMs) present significant challenges for their efficient deployment. Recently, there is an increasing interest in quantizing weights to extremely low precision while offsetting the resulting error with low-rank, high-precision error reconstruction terms. The combination of quantization and low-rank approximation is now popular in both adapter-based, parameter-efficient fine-tuning methods such as LoftQ and low-precision inference techniques including ZeroQuant-V2. Usually, the low-rank terms are calculated via the singular value decomposition (SVD) of the weight quantization error, minimizing the Frobenius and spectral norms of the weight approximation error. Recent methods like LQ-LoRA and LQER introduced hand-crafted heuristics to minimize errors in layer outputs (activations) rather than weights, resulting improved quantization results. However, these heuristic methods lack an analytical solution to guide the design of quantization error reconstruction terms. In this paper, we revisit this problem and formulate an analytical framework, named Quantization Error Reconstruction Analysis (QERA), and offer a closed-form solution to the problem. We show QERA benefits both existing low-precision fine-tuning and inference methods -- QERA achieves a fine-tuned accuracy gain of $\Delta_{\text{acc}}$ = 6.05% of 2-bit RoBERTa-base on GLUE compared to LoftQ; and obtains $\Delta_{\text{acc}}$ = 2.97% higher post-training quantization accuracy of 4-bit Llama-3.1-70B on average than ZeroQuant-V2 and $\Delta_{\text{ppl}}$ = - 0.28 lower perplexity on WikiText2 than LQER.

Wireless networks supporting artificial intelligence have gained significant attention, with Over-the-Air Federated Learning emerging as a key application due to its unique transmission and distributed computing characteristics. This paper derives error bounds for Over-the-Air Federated Learning in a Cell-free MIMO system and formulates an optimization problem to minimize optimality gap via joint optimization of power control and beamforming. We introduce the MOP-LOFPC algorithm, which employs Lyapunov optimization to decouple long-term constraints across rounds while requiring only causal channel state information. Experimental results demonstrate that MOP-LOFPC achieves a better and more flexible trade-off between the model's training loss and adherence to long-term power constraints compared to existing baselines.

We introduce FabricDiffusion, a method for transferring fabric textures from a single clothing image to 3D garments of arbitrary shapes. Existing approaches typically synthesize textures on the garment surface through 2D-to-3D texture mapping or depth-aware inpainting via generative models. Unfortunately, these methods often struggle to capture and preserve texture details, particularly due to challenging occlusions, distortions, or poses in the input image. Inspired by the observation that in the fashion industry, most garments are constructed by stitching sewing patterns with flat, repeatable textures, we cast the task of clothing texture transfer as extracting distortion-free, tileable texture materials that are subsequently mapped onto the UV space of the garment. Building upon this insight, we train a denoising diffusion model with a large-scale synthetic dataset to rectify distortions in the input texture image. This process yields a flat texture map that enables a tight coupling with existing Physically-Based Rendering (PBR) material generation pipelines, allowing for realistic relighting of the garment under various lighting conditions. We show that FabricDiffusion can transfer various features from a single clothing image including texture patterns, material properties, and detailed prints and logos. Extensive experiments demonstrate that our model significantly outperforms state-to-the-art methods on both synthetic data and real-world, in-the-wild clothing images while generalizing to unseen textures and garment shapes., Comment: Accepted to SIGGRAPH Asia 2024. Project page: https://humansensinglab.github.io/fabric-diffusion

Existing unified methods typically treat multi-degradation image restoration as a multi-task learning problem. Despite performing effectively compared to single degradation restoration methods, they overlook the utilization of commonalities and specificities within multi-task restoration, thereby impeding the model's performance. Inspired by the success of deep generative models and fine-tuning techniques, we proposed a universal image restoration framework based on multiple low-rank adapters (LoRA) from multi-domain transfer learning. Our framework leverages the pre-trained generative model as the shared component for multi-degradation restoration and transfers it to specific degradation image restoration tasks using low-rank adaptation. Additionally, we introduce a LoRA composing strategy based on the degradation similarity, which adaptively combines trained LoRAs and enables our model to be applicable for mixed degradation restoration. Extensive experiments on multiple and mixed degradations demonstrate that the proposed universal image restoration method not only achieves higher fidelity and perceptual image quality but also has better generalization ability than other unified image restoration models. Our code is available at https://github.com/Justones/UIR-LoRA.

Augmented Reality (AR) and Virtual Reality (VR), require miniaturized displays with ultrahigh pixel densities. Here, we demonstrate an individually addressable subwavelength OLED pixel based on a nanoscale electrode capable of supporting plasmonic modes. Our approach is based on the notion that when scaling down pixel size, the 2D planar geometry of conventional organic light-emitting diodes (OLEDs) evolves into a significantly more complex 3D geometry governed by sharp nanoelectrode contours. These cause (i) spatially imbalanced charge carrier transport and recombination, resulting in a low quantum efficiency, and (ii) filament growth, leading to rapid device failure. Here, we circumvent such effects by selectively covering sharp electrode contours with an insulating layer, while utilizing a nano-aperture in flat areas of the electrode. We thereby ensure controlled charge carrier injection and recombination at the nanoscale and suppress filament growth. As a proof of principle, we first demonstrate stable and efficient hole injection from Au nanoelectrodes in hole-only devices with above 90 % pixel yield and longtime operation stability and then a complete vertical OLED pixel with an individually addressable nanoelectrode (300 x 300 nm$^{2}$), highlighting the potential to further leverage plasmonic nanoantenna effects to enhance the performance and functionality of nano-OLEDs.

Hierarchical search in millimeter-wave (mmWave) communications incurs significant beam training overhead and delay, especially in a dynamic environment. Deep learning-enabled beam prediction is promising to significantly mitigate the overhead and delay, efficiently utilizing the site-specific channel prior. In this work, we propose to jointly optimize a data- and model-driven probe beam module and a cascaded data-driven beam predictor, with limitations in that the probe and communicate beams are restricted within the manifold space of uniform planer array and quantization of the phase modulator. First, The probe beam module senses the mmWave channel with a complex-valued neural network and outputs the counterpart RSRPs of probe beams. Second, the beam predictor estimates the RSRPs in the entire beamspace to minimize the prediction cross entropy and selects the optimal beam with the maximum RSRP value for data transmission. Additionally, we propose to add noise to the phase variables in the probe beam module, against quantization error. Simulation results show the effectiveness of our proposed scheme.

The widespread success of deep learning models today is owed to the curation of extensive datasets significant in size and complexity. However, such models frequently pick up inherent biases in the data during the training process, leading to unreliable predictions. Diagnosing and debiasing datasets is thus a necessity to ensure reliable model performance. In this paper, we present CONBIAS, a novel framework for diagnosing and mitigating Concept co-occurrence Biases in visual datasets. CONBIAS represents visual datasets as knowledge graphs of concepts, enabling meticulous analysis of spurious concept co-occurrences to uncover concept imbalances across the whole dataset. Moreover, we show that by employing a novel clique-based concept balancing strategy, we can mitigate these imbalances, leading to enhanced performance on downstream tasks. Extensive experiments show that data augmentation based on a balanced concept distribution augmented by CONBIAS improves generalization performance across multiple datasets compared to state-of-the-art methods. We will make our code and data publicly available.

The BICEP3 and BICEP Array polarimeters are small-aperture refracting telescopes located at the South Pole designed to measure primordial gravitational wave signatures in the Cosmic Microwave Background (CMB) polarization, predicted by inflation. Constraining the inflationary signal requires not only excellent sensitivity, but also careful control of instrumental systematics. Both instruments use antenna-coupled orthogonally polarized detector pairs, and the polarized sky signal is reconstructed by taking the difference in each detector pair. As a result, the differential response between detectors within a pair becomes an important systematic effect we must control. Additionally, mapping the intensity and polarization response in regions away from the main beam can inform how sidelobe levels affect CMB measurements. Extensive calibration measurements are taken in situ every austral summer for control of instrumental systematics and instrument characterisation. In this work, we detail the set of beam calibration measurements that we conduct on the BICEP receivers, from deep measurements of main beam response to polarized beam response and sidelobe mapping. We discuss the impact of these measurements for instrumental systematics studies and design choices for future CMB receivers., Comment: 13 pages, 7 figures, 1 table, Proceedings paper SPIE 2024

Probing beam measurement (PBM)-based hybrid beamforming provides a feasible solution for cell-free MIMO. In this letter, we propose a novel probing beam optimization framework where three collaborative modules respectively realize PBM augmentation, sum-rate prediction and probing beam optimization. Specifically, the PBM augmentation model integrates the conditional variational auto-encoder (CVAE) and mixture density networks and adopts correlated PBM distribution with full-covariance, for which a Cholesky-decomposition based training is introduced to address the issues of covariance legality and numerical stability. Simulations verify the better performance of the proposed augmentation model compared to the traditional CVAE and the efficiency of proposed optimization framework.

In this letter, the maximization of resistible jamming power is studied for multi-user downlink millimeter-wave cell-free multiple-input-multiple-output (CF-MIMO) systems. We propose an alternate optimization-based anti-jamming hybrid beamforming (AO-AJHBF) scheme. For receiving beamforming, more practical prior about the jamming channel, i.e., second-order statistics rather than instantaneous information, is exploited via maximizing the generalized Rayleigh quotient. For transmitting beamforming, we use the max-min fairness principle and propose a low-complexity projected gradient ascent-based method to circumvent the excessive computation of semi-definite relaxation (SDR). Simulations verify the performance advantage of proposed AO-AJHBF over schemes based on weighted minimum mean square error and SDR methods.

Probability estimation of tree topologies is one of the fundamental tasks in phylogenetic inference. The recently proposed subsplit Bayesian networks (SBNs) provide a powerful probabilistic graphical model for tree topology probability estimation by properly leveraging the hierarchical structure of phylogenetic trees. However, the expectation maximization (EM) method currently used for learning SBN parameters does not scale up to large data sets. In this paper, we introduce several computationally efficient methods for training SBNs and show that variance reduction could be the key for better performance. Furthermore, we also introduce the variance reduction technique to improve the optimization of SBN parameters for variational Bayesian phylogenetic inference (VBPI). Extensive synthetic and real data experiments demonstrate that our methods outperform previous baseline methods on the tasks of tree topology probability estimation as well as Bayesian phylogenetic inference using SBNs., Comment: 23 pages, 6 figures, 7 tables

We prove Weyl laws for Schr\"odinger operators with critically singular potentials on compact manifolds with boundary. We also improve the Weyl remainder estimates under the condition that the set of all periodic geodesic billiards has measure 0. These extend the classical results by Seeley, Ivrii and Melrose. The proof uses the Gaussian heat kernel bounds for short times and a perturbation argument involving the wave equation., Comment: 19 pages

The Varying Speed of Light (VSL) model describes how the speed of light in a vacuum changes with cosmological redshift. Despite numerous models, there is little observational evidence for this variation. While the speed of light can be accurately measured by physical means, cosmological methods are rarely used. Previous studies quantified the speed of light at specific redshifts using Gaussian processes and reconstructed the redshift-dependent function $c(z)$. It is crucial to quantify the speed of light across varying redshifts. We use the latest data on angular diameter distances $D_A(z)$ and Hubble parameters $H(z)$ from baryon acoustic oscillation (BAO) and cosmic chronometer measurements in the redshift interval $z\in[0.07,1.965]$. The speed of light $c(z)$ is determined using Gaussian and deep Gaussian processes to reconstruct $H(z)$, $D_A(z)$, and $D^{\prime}_A(z)$. Furthermore, we conduct comparisons across three distinct models, encompassing two renowned VSL models. We get the result of the parameters constraints in the models (1) for the ``$c$-c" model, $c_0=29492.6 \pm^{6.2}_{5.3} \mathrm{~km} \mathrm{~s}^{-1}$. (2) For the ``$c$-cl" model, $c_0=29665.5 \pm^{11.2}_{11.4}\mathrm{~km} \mathrm{~s}^{-1}$ and $n=0.05535 \pm^{0.00008}_{0.00007}$. (3) For the ``$c$-CPL" model, $c_0=29555.7 \pm^{13.3}_{13.2} \mathrm{~km} \mathrm{~s}^{-1}$ and $n=-0.0607 \pm 0.0001$. Based on our findings, it may be inferred that Barrow's classical VSL model is not a suitable fit for our data. In contrast, the widely recognized Chevallier-Polarski-Linder (CPL) VSL model, under some circumstances, as well as the universal ``c is constant" model, demonstrate a satisfactory ability to account for our findings., Comment: 15 pages, 9 figures; Accepted for publication in MNRAS

Slow inference speed is one of the most crucial concerns for deploying multi-view 3D detectors to tasks with high real-time requirements like autonomous driving. Although many sparse query-based methods have already attempted to improve the efficiency of 3D detectors, they neglect to consider the backbone, especially when using Vision Transformers (ViT) for better performance. To tackle this problem, we explore the efficient ViT backbones for multi-view 3D detection via token compression and propose a simple yet effective method called TokenCompression3D (ToC3D). By leveraging history object queries as foreground priors of high quality, modeling 3D motion information in them, and interacting them with image tokens through the attention mechanism, ToC3D can effectively determine the magnitude of information densities of image tokens and segment the salient foreground tokens. With the introduced dynamic router design, ToC3D can weigh more computing resources to important foreground tokens while compressing the information loss, leading to a more efficient ViT-based multi-view 3D detector. Extensive results on the large-scale nuScenes dataset show that our method can nearly maintain the performance of recent SOTA with up to 30% inference speedup, and the improvements are consistent after scaling up the ViT and input resolution. The code will be made at https://github.com/DYZhang09/ToC3D., Comment: Accepted by ECCV 2024

Reconstructing the evolutionary history relating a collection of molecular sequences is the main subject of modern Bayesian phylogenetic inference. However, the commonly used Markov chain Monte Carlo methods can be inefficient due to the complicated space of phylogenetic trees, especially when the number of sequences is large. An alternative approach is variational Bayesian phylogenetic inference (VBPI) which transforms the inference problem into an optimization problem. While effective, the default diagonal lognormal approximation for the branch lengths of the tree used in VBPI is often insufficient to capture the complexity of the exact posterior. In this work, we propose a more flexible family of branch length variational posteriors based on semi-implicit hierarchical distributions using graph neural networks. We show that this semi-implicit construction emits straightforward permutation equivariant distributions, and therefore can handle the non-Euclidean branch length space across different tree topologies with ease. To deal with the intractable marginal probability of semi-implicit variational distributions, we develop several alternative lower bounds for stochastic optimization. We demonstrate the effectiveness of our proposed method over baseline methods on benchmark data examples, in terms of both marginal likelihood estimation and branch length posterior approximation., Comment: 26 pages, 7 figures

Burq-G\'erard-Tzvetkov and Hu established $L^p$ estimates for the restriction of Laplace-Beltrami eigenfunctions to submanifolds. We investigate the eigenfunctions of the Schr\"odinger operators with critically singular potentials, and estimate the $L^p$ norms and period integrals for their restriction to submanifolds. Recently, Blair-Sire-Sogge obtained global $L^p$ bounds for Schr\"odinger eigenfunctions by the resolvent method. Due to the Sobolev trace inequalities, the resolvent method cannot work for submanifolds of all dimensions. We get around this difficulty and establish spectral projection bounds by the wave kernel techniques and the bootstrap argument involving an induction on the dimensions of the submanifolds., Comment: 36 pages, 1 figure

The Perspective-n-Point (PnP) problem has been widely studied in the literature and applied in various vision-based pose estimation scenarios. However, existing methods ignore the anisotropy uncertainty of observations, as demonstrated in several real-world datasets in this paper. This oversight may lead to suboptimal and inaccurate estimation, particularly in the presence of noisy observations. To this end, we propose a generalized maximum likelihood PnP solver, named GMLPnP, that minimizes the determinant criterion by iterating the GLS procedure to estimate the pose and uncertainty simultaneously. Further, the proposed method is decoupled from the camera model. Results of synthetic and real experiments show that our method achieves better accuracy in common pose estimation scenarios, GMLPnP improves rotation/translation accuracy by 4.7%/2.0% on TUM-RGBD and 18.6%/18.4% on KITTI-360 dataset compared to the best baseline. It is more accurate under very noisy observations in a vision-based UAV localization task, outperforming the best baseline by 34.4% in translation estimation accuracy.

Recent theoretical efforts predicted a type of unconventional antiferromagnet characterized by the crystal symmetry C (rotation or mirror), which connects antiferromagnetic sublattices in real space and simultaneously couples spin and momentum in reciprocal space. This results in a unique C-paired spin-valley locking (SVL) and corresponding novel properties such as piezomagnetism and noncollinear spin current even without spin-orbit coupling. However, the unconventional antiferromagnets reported thus far are not layered materials, limiting their potential in spintronic applications. Additionally, they do not meet the necessary symmetry requirements for nonrelativistic spin current. Here, we report the realization of C-paired SVL in a layered room-temperature antiferromagnetic compound, Rb1-{\delta}V2Te2O. Spin resolved photoemission measurements directly demonstrate the opposite spin splitting between C-paired valleys. Quasi-particle interference patterns reveal the suppression of inter-valley scattering due to the spin selection rules, as a direct consequence of C-paired SVL. All these experiments are well consistent with the results obtained from first-principles calculations. Our observations represent the first realization of layered antiferromagnets with C-paired SVL, enabling both the advantages of layered materials and possible control through crystal symmetry manipulation. These results hold significant promise and broad implications for advancements in magnetism, electronics, and information technology., Comment: 22 pages, 5 figures

Recent progress in neural rendering has brought forth pioneering methods, such as NeRF and Gaussian Splatting, which revolutionize view rendering across various domains like AR/VR, gaming, and content creation. While these methods excel at interpolating {\em within the training data}, the challenge of generalizing to new scenes and objects from very sparse views persists. Specifically, modeling 3D humans from sparse views presents formidable hurdles due to the inherent complexity of human geometry, resulting in inaccurate reconstructions of geometry and textures. To tackle this challenge, this paper leverages recent advancements in Gaussian Splatting and introduces a new method to learn generalizable human Gaussians that allows photorealistic and accurate view-rendering of a new human subject from a limited set of sparse views in a feed-forward manner. A pivotal innovation of our approach involves reformulating the learning of 3D Gaussian parameters into a regression process defined on the 2D UV space of a human template, which allows leveraging the strong geometry prior and the advantages of 2D convolutions. In addition, a multi-scaffold is proposed to effectively represent the offset details. Our method outperforms recent methods on both within-dataset generalization as well as cross-dataset generalization settings.

The recent advancements in text-to-image generative models have been remarkable. Yet, the field suffers from a lack of evaluation metrics that accurately reflect the performance of these models, particularly lacking fine-grained metrics that can guide the optimization of the models. In this paper, we propose EvalAlign, a metric characterized by its accuracy, stability, and fine granularity. Our approach leverages the capabilities of Multimodal Large Language Models (MLLMs) pre-trained on extensive data. We develop evaluation protocols that focus on two key dimensions: image faithfulness and text-image alignment. Each protocol comprises a set of detailed, fine-grained instructions linked to specific scoring options, enabling precise manual scoring of the generated images. We supervised fine-tune (SFT) the MLLM to align with human evaluative judgments, resulting in a robust evaluation model. Our evaluation across 24 text-to-image generation models demonstrate that EvalAlign not only provides superior metric stability but also aligns more closely with human preferences than existing metrics, confirming its effectiveness and utility in model assessment., Comment: Project page: https://sais-fuxi.github.io/projects/evalalign/

Low-rank Adaption (LoRA) has been the de-facto parameter-efficient fine-tuning technique for large language models. We present HeteroLoRA, a light-weight search algorithm that leverages zero-cost proxies to allocate the limited LoRA trainable parameters across the model for better fine-tuned performance. In addition to the allocation for the standard LoRA-adapted models, we also demonstrate the efficacy of HeteroLoRA by performing the allocation in a more challenging search space that includes LoRA modules and LoRA-adapted shortcut connections. Experiments show that HeteroLoRA enables improvements in model performance given the same parameter budge. For example, on MRPC, we see an improvement of 1.6% in accuracy with similar training parameter budget. We will open-source our algorithm once the paper is accepted., Comment: Accepted at ICML2024 ES-FoMo-II Workshop

Grouped-query attention (GQA) has been widely adopted in LLMs to mitigate the complexity of multi-head attention (MHA). To transform an MHA to a GQA, neighbour queries in MHA are evenly split into groups where each group shares the value and key layers. In this work, we propose AsymGQA, an activation-informed approach to asymmetrically grouping an MHA to a GQA for better model performance. Our AsymGQA outperforms the GQA within the same model size budget. For example, AsymGQA LLaMA-2-7B has an accuracy increase of 7.5% on MMLU compared to neighbour grouping. Our approach addresses the GQA's trade-off problem between model performance and hardware efficiency., Comment: Accepted at ICML2024 ES-FoMo-II Workshop

Point forecasting in univariate random walks is an important yet challenging research topic. Many attempts at this task often fail to surpass the na\"ive baseline because of the randomness of the data and the improper utilization of exogenous variables as features. In view of the limitations of existing random walk forecasting methods, this study introduces a variant definition of random walks, proposing that point forecasting can be improved beyond the na\"ive baseline through the fusion of movement and na\"ive predictions (FMNP). FMNP naturally bridges movement prediction and point forecasting. It employs an exogenous variable to provide a consistent movement prediction for the target variable and uses a linear regression to combine movement and na\"ive predictions. In forecasting five financial time series in the U.S. market with the FTSE opening price as the exogenous variable, FMNP consistently outperforms na\"ive baselines and is superior to baseline models such as ARIMA, MA, MLP, DNN, LSTM, and CNN-LSTM. FMNP is particularly advantageous when accurate point predictions are challenging but accurate movement predictions are attainable, translating movement predictions into point forecasts in random walk contexts.

Seams are areas of overlapping fabric formed by stitching two or more pieces of fabric together in the cut-and-sew apparel manufacturing process. In SeamPose, we repurposed seams as capacitive sensors in a shirt for continuous upper-body pose estimation. Compared to previous all-textile motion-capturing garments that place the electrodes on the clothing surface, our solution leverages existing seams inside of a shirt by machine-sewing insulated conductive threads over the seams. The unique invisibilities and placements of the seams afford the sensing shirt to look and wear similarly as a conventional shirt while providing exciting pose-tracking capabilities. To validate this approach, we implemented a proof-of-concept untethered shirt with 8 capacitive sensing seams. With a 12-participant user study, our customized deep-learning pipeline accurately estimates the relative (to the pelvis) upper-body 3D joint positions with a mean per joint position error (MPJPE) of 6.0 cm. SeamPose represents a step towards unobtrusive integration of smart clothing for everyday pose estimation.

Self-recording eating behaviors is a step towards a healthy lifestyle recommended by many health professionals. However, the current practice of manually recording eating activities using paper records or smartphone apps is often unsustainable and inaccurate. Smart glasses have emerged as a promising wearable form factor for tracking eating behaviors, but existing systems primarily identify when eating occurs without capturing details of the eating activities (E.g., what is being eaten). In this paper, we present EchoGuide, an application and system pipeline that leverages low-power active acoustic sensing to guide head-mounted cameras to capture egocentric videos, enabling efficient and detailed analysis of eating activities. By combining active acoustic sensing for eating detection with video captioning models and large-scale language models for retrieval augmentation, EchoGuide intelligently clips and analyzes videos to create concise, relevant activity records on eating. We evaluated EchoGuide with 9 participants in naturalistic settings involving eating activities, demonstrating high-quality summarization and significant reductions in video data needed, paving the way for practical, scalable eating activity tracking., Comment: Accepted at ISWC '24

Tabular synthesis models remain ineffective at capturing complex dependencies, and the quality of synthetic data is still insufficient for comprehensive downstream tasks, such as prediction under distribution shifts, automated decision-making, and cross-table understanding. A major challenge is the lack of prior knowledge about underlying structures and high-order relationships in tabular data. We argue that a systematic evaluation on high-order structural information for tabular data synthesis is the first step towards solving the problem. In this paper, we introduce high-order structural causal information as natural prior knowledge and provide a benchmark framework for the evaluation of tabular synthesis models. The framework allows us to generate benchmark datasets with a flexible range of data generation processes and to train tabular synthesis models using these datasets for further evaluation. We propose multiple benchmark tasks, high-order metrics, and causal inference tasks as downstream tasks for evaluating the quality of synthetic data generated by the trained models. Our experiments demonstrate to leverage the benchmark framework for evaluating the model capability of capturing high-order structural causal information. Furthermore, our benchmarking results provide an initial assessment of state-of-the-art tabular synthesis models. They have clearly revealed significant gaps between ideal and actual performance and how baseline methods differ. Our benchmark framework is available at URL https://github.com/TURuibo/CauTabBench.

Smart glasses have become more prevalent as they provide an increasing number of applications for users. They store various types of private information or can access it via connections established with other devices. Therefore, there is a growing need for user identification on smart glasses. In this paper, we introduce a low-power and minimally-obtrusive system called SonicID, designed to authenticate users on glasses. SonicID extracts unique biometric information from users by scanning their faces with ultrasonic waves and utilizes this information to distinguish between different users, powered by a customized binary classifier with the ResNet-18 architecture. SonicID can authenticate users by scanning their face for 0.06 seconds. A user study involving 40 participants confirms that SonicID achieves a true positive rate of 97.4%, a false positive rate of 4.3%, and a balanced accuracy of 96.6% using just 1 minute of training data collected for each new user. This performance is relatively consistent across different remounting sessions and days. Given this promising performance, we further discuss the potential applications of SonicID and methods to improve its performance in the future., Comment: 27 pages, 6 tables, 9 figures

We investigate proton-antiproton ($p\bar{p}$) pair production via photon-photon fusion in the ultra-peripheral collisions at RHIC, employing a joint impact parameter and transverse momentum dependent formalism. We consider proton exchange, $s$-channel resonance and hand-bag mechanisms, predicting differential distributions of $p\bar p$ production. Our theoretical predictions can be tested against future measurements at RHIC, to enhance our understanding of photon-photon interactions in strong electromagnetic fields., Comment: 6 pages, 4 figures

Physical neural networks (PNNs) are a class of neural-like networks that leverage the properties of physical systems to perform computation. While PNNs are so far a niche research area with small-scale laboratory demonstrations, they are arguably one of the most underappreciated important opportunities in modern AI. Could we train AI models 1000x larger than current ones? Could we do this and also have them perform inference locally and privately on edge devices, such as smartphones or sensors? Research over the past few years has shown that the answer to all these questions is likely "yes, with enough research": PNNs could one day radically change what is possible and practical for AI systems. To do this will however require rethinking both how AI models work, and how they are trained - primarily by considering the problems through the constraints of the underlying hardware physics. To train PNNs at large scale, many methods including backpropagation-based and backpropagation-free approaches are now being explored. These methods have various trade-offs, and so far no method has been shown to scale to the same scale and performance as the backpropagation algorithm widely used in deep learning today. However, this is rapidly changing, and a diverse ecosystem of training techniques provides clues for how PNNs may one day be utilized to create both more efficient realizations of current-scale AI models, and to enable unprecedented-scale models., Comment: 29 pages, 4 figures

Deep Neural Networks (DNNs) excel in learning hierarchical representations from raw data, such as images, audio, and text. To compute these DNN models with high performance and energy efficiency, these models are usually deployed onto customized hardware accelerators. Among various accelerator designs, dataflow architecture has shown promising performance due to its layer-pipelined structure and its scalability in data parallelism. Exploiting weights and activations sparsity can further enhance memory storage and computation efficiency. However, existing approaches focus on exploiting sparsity in non-dataflow accelerators, which cannot be applied onto dataflow accelerators because of the large hardware design space introduced. As such, this could miss opportunities to find an optimal combination of sparsity features and hardware designs. In this paper, we propose a novel approach to exploit unstructured weights and activations sparsity for dataflow accelerators, using software and hardware co-optimization. We propose a Hardware-Aware Sparsity Search (HASS) to systematically determine an efficient sparsity solution for dataflow accelerators. Over a set of models, we achieve an efficiency improvement ranging from 1.3$\times$ to 4.2$\times$ compared to existing sparse designs, which are either non-dataflow or non-hardware-aware. Particularly, the throughput of MobileNetV3 can be optimized to 4895 images per second. HASS is open-source: \url{https://github.com/Yu-Zhewen/HASS}, Comment: accepted to FPL2024

Multi-modal salient object detection (MSOD) aims to boost saliency detection performance by integrating visible sources with depth or thermal infrared ones. Existing methods generally design different fusion schemes to handle certain issues or challenges. Although these fusion schemes are effective at addressing specific issues or challenges, they may struggle to handle multiple complex challenges simultaneously. To solve this problem, we propose a novel adaptive fusion bank that makes full use of the complementary benefits from a set of basic fusion schemes to handle different challenges simultaneously for robust MSOD. We focus on handling five major challenges in MSOD, namely center bias, scale variation, image clutter, low illumination, and thermal crossover or depth ambiguity. The fusion bank proposed consists of five representative fusion schemes, which are specifically designed based on the characteristics of each challenge, respectively. The bank is scalable, and more fusion schemes could be incorporated into the bank for more challenges. To adaptively select the appropriate fusion scheme for multi-modal input, we introduce an adaptive ensemble module that forms the adaptive fusion bank, which is embedded into hierarchical layers for sufficient fusion of different source data. Moreover, we design an indirect interactive guidance module to accurately detect salient hollow objects via the skip integration of high-level semantic information and low-level spatial details. Extensive experiments on three RGBT datasets and seven RGBD datasets demonstrate that the proposed method achieves the outstanding performance compared to the state-of-the-art methods. The code and results are available at https://github.com/Angknpng/LAFB., Comment: Accepted by TCSVT 2024

We introduce MunchSonic, an AI-powered active acoustic sensing system integrated into eyeglasses to track fine-grained dietary actions. MunchSonic emits inaudible ultrasonic waves from the eyeglass frame, with the reflected signals capturing detailed positions and movements of body parts, including the mouth, jaw, arms, and hands involved in eating. These signals are processed by a deep learning pipeline to classify six actions: hand-to-mouth movements for food intake, chewing, drinking, talking, face-hand touching, and other activities (null). In an unconstrained study with 12 participants, MunchSonic achieved a 93.5% macro F1-score in a user-independent evaluation with a 2-second resolution in tracking these actions, also demonstrating its effectiveness in tracking eating episodes and food intake frequency within those episodes., Comment: 8 pages, 7 figures

Modern Machine Learning models are expensive IP and business competitiveness often depends on keeping this IP confidential. This in turn restricts how these models are deployed -- for example it is unclear how to deploy a model on-device without inevitably leaking the underlying model. At the same time, confidential computing technologies such as Multi-Party Computation or Homomorphic encryption remain impractical for wide adoption. In this paper we take a different approach and investigate feasibility of ML-specific mechanisms that deter unauthorized model use by restricting the model to only be usable on specific hardware, making adoption on unauthorized hardware inconvenient. That way, even if IP is compromised, it cannot be trivially used without specialised hardware or major model adjustment. In a sense, we seek to enable cheap locking of machine learning models into specific hardware. We demonstrate that locking mechanisms are feasible by either targeting efficiency of model representations, such making models incompatible with quantisation, or tie the model's operation on specific characteristics of hardware, such as number of cycles for arithmetic operations. We demonstrate that locking comes with negligible work and latency overheads, while significantly restricting usability of the resultant model on unauthorized hardware., Comment: 10 pages, 2 figures of main text; 14 pages, 16 figures of appendices

Semi-implicit variational inference (SIVI) extends traditional variational families with semi-implicit distributions defined in a hierarchical manner. Due to the intractable densities of semi-implicit distributions, classical SIVI often resorts to surrogates of evidence lower bound (ELBO) that would introduce biases for training. A recent advancement in SIVI, named SIVI-SM, utilizes an alternative score matching objective made tractable via a minimax formulation, albeit requiring an additional lower-level optimization. In this paper, we propose kernel SIVI (KSIVI), a variant of SIVI-SM that eliminates the need for lower-level optimization through kernel tricks. Specifically, we show that when optimizing over a reproducing kernel Hilbert space (RKHS), the lower-level problem has an explicit solution. This way, the upper-level objective becomes the kernel Stein discrepancy (KSD), which is readily computable for stochastic gradient descent due to the hierarchical structure of semi-implicit variational distributions. An upper bound for the variance of the Monte Carlo gradient estimators of the KSD objective is derived, which allows us to establish novel convergence guarantees of KSIVI. We demonstrate the effectiveness and efficiency of KSIVI on both synthetic distributions and a variety of real data Bayesian inference tasks., Comment: ICML 2024 camera ready

Measurement-device-independent quantum secret sharing (MDI-QSS) can eliminate all the security loopholes associated with imperfect measurement devices and greatly enhance QS's security under practical experimental condition. MDI-QSS requires each communication user to send single photon to the measurement party for the coincident measurement. However, the unsynchronization of the transmitted photons greatly limits MDI-QSS's practical performance.In the paper, we propose a high-efficient quantum memory (QM)-assisted MDI-QSS protocol, which employs the QM-assisted synchronization of three heralded single-photon sources to efficiently generate three simultaneous single-photon states. The QM constructed with all-optical, polarization-insensitive storage loop has superior performance in terms of bandwidth, storage efficiency, and noise resistance, and is feasible under current experiment conditions. Combining with the decoy-state method, we perform the numerical simulation of the secure key rate in the symmetric model without considering the finite-size effect. The simulation results show that our QM-assisted MDI-QSS protocol exhibit largely improved secure key rate and maximal photon transmission distance compared with all existing MDI-QSS protocols without QM. Our protocol provides a promising way for implementing the high-efficient long-distance MDI-QSS in the near future., Comment: 11 pages, 6 figures