Search

Data preprocessing pipelines, which includes data decoding, cleaning, and transforming, are a crucial component of Machine Learning (ML) training. Thy are computationally intensive and often become a major bottleneck, due to the increasing performance gap between the CPUs used for preprocessing and the GPUs used for model training. Recent studies show that a significant number of CPUs across several machines are required to achieve sufficient throughput to saturate the GPUs, leading to increased resource and energy consumption. When the pipeline involves vocabulary generation, the preprocessing performance scales poorly due to significant row-wise synchronization overhead between different CPU cores and servers. To address this limitation, in this paper we present the design of Piper, a hardware accelerator for tabular data preprocessing, prototype it on FPGAs, and demonstrate its potential for training pipelines of commercial recommender systems. Piper achieves 4.7 $\sim$ 71.3$\times$ speedup in latency over a 128-core CPU server and outperforms a data-center GPU by 4.8$\sim$ 20.3$\times$ when using binary input. The impressive performance showcases Piper's potential to increase the efficiency of data preprocessing pipelines and significantly reduce their resource consumption., Comment: 12 pages, 10 figures, 4 tables

Over 700 bright millisecond-duration radio transients, known as Fast Radio Bursts (FRBs), have been identified to date. Nevertheless, the origin of FRBs remains unknown. The two repeating FRBs (FRB 20121102A and FRB 20190520B) have been verified to be associated with persistent radio sources (PRSs), making them the best candidates to study the nature of FRBs. Monitoring the variability in PRSs is essential for understanding their physical nature. We conducted 22 observations of the PRSs linked to FRB 20121102A and FRB 20190520B using the Karl G. Jansky Very Large Array (VLA), to study their variability. We have observed significant flux variability for the PRSs of FRB 20121102A and FRB 20190520B, with a confidence level exceeding 99.99%, based on the observations covering the longest timescale recorded to date. The observed variability of the two PRSs exhibits no significant difference in amplitude across both short and long timescales. We found that the radio-derived star formation rates of the two FRB hosts are significantly higher than those measured by the optical $H_{\alpha}$ emissions, indicating that their host galaxies are highly obscured or most radio emissions are not from star formation processes. The observed timescale of PRS flux evolution constrained the magnetic field of FRB 20121102A with $B_\parallel\gtrsim1~{\rm mG}$ and FRB 20190520B with $B_\parallel\gtrsim0.1~{\rm mG}$., Comment: 16 pages, 7 figures, accepted by ApJ

The spin filtering effect, distinct decaying lengths experienced by oppositely spin-polarized electrons in a magnetic barrier, generally occurs in ferromagnetic (FM) insulators or semiconductors. With the rise of altermagnetic (ALM) materials which exhibit similar capability of spin-polarizing electrons with ferromagnets, it is a nature question whether the ALM insulators or semiconductors can also act as unique barriers for the spin splitting effect. Here, through first-principles calculations, we investigated the complex band structure of the ALM insulator FeF$_2$ and found that it possesses an anisotropic spin filtering effect: along the [001] direction of FeF$_2$, a current remains spin-neutral but has locally nonvanishing spin polarizations in the momentum space; moreover, along the [110] direction of FeF$_2$, a current will be globally spin-polarized by different attenuation lengths of oppositely spin-polarized electrons. Leveraging this anisotropic spin filtering effect, we designed two types of MTJs with the ALM barrier: ALM electrode/ALM insulator barrier/non-magnetic (NM) electrode and FM electrode/ALM insulator barrier/NM electrode, using RuO$_2$(001)/FeF$_2$/IrO$_2$ and CrO$_2$(110)/FeF$_2$/IrO$_2$ as the corresponding prototypes, respectively. We found that these two proposed MTJs exhibited the tunneling magnetoresistance (TMR) ratios of 216\% and 3956\%, by matching the conduction channels of the electrodes and the spin-resolved lowest decay rate of the barrier in the momentum space. Our work deepens and generalizes understanding toward the spin filtering effect for the rising ALM insulators and semiconductors, and broadens applications of the AFM spintronics., Comment: 6 pages, 4 figures

Intelligent reflecting surfaces (IRSs) have been regarded as a promising enabler for future wireless communication systems. In the literature, IRSs have been considered power-free or assumed to have constant power consumption. However, recent experimental results have shown that for positive-intrinsic-negative (PIN) diode-based IRSs, the power consumption dynamically changes with the phase shift configuration. This phase shift-dependent power consumption (PS-DPC) introduces a challenging power allocation problem between base station (BS) and IRS. To tackle this issue, in this paper, we investigate a rate maximization problem for IRS-assisted systems under a practical PS-DPC model. For the single-user case, we propose a generalized Benders decomposition-based beamforming method to maximize the achievable rate while satisfying a total system power consumption constraint. Moreover, we propose a low-complexity beamforming design, where the powers allocated to BS and IRS are optimized offline based on statistical channel state information. Furthermore, for the multi-user case, we solve an equivalent weighted mean square error minimization problem with two different joint power allocation and phase shift optimization methods. Simulation results indicate that compared to baseline schemes, our proposed methods can flexibly optimize the power allocation between BS and IRS, thus achieving better performance. The optimized power allocation strategy strongly depends on the system power budget. When the system power budget is high, the PS-DPC is not the dominant factor in the system power consumption, allowing the IRS to turn on as many PIN diodes as needed to achieve high beamforming quality. When the system power budget is limited, however, more power tends to be allocated to the BS to enhance the transmit power, resulting in a lower beamforming quality at the IRS due to the reduced PS-DPC budget.

Understanding how biological visual systems process information is challenging because of the nonlinear relationship between visual input and neuronal responses. Artificial neural networks allow computational neuroscientists to create predictive models that connect biological and machine vision. Machine learning has benefited tremendously from benchmarks that compare different model on the same task under standardized conditions. However, there was no standardized benchmark to identify state-of-the-art dynamic models of the mouse visual system. To address this gap, we established the Sensorium 2023 Benchmark Competition with dynamic input, featuring a new large-scale dataset from the primary visual cortex of ten mice. This dataset includes responses from 78,853 neurons to 2 hours of dynamic stimuli per neuron, together with the behavioral measurements such as running speed, pupil dilation, and eye movements. The competition ranked models in two tracks based on predictive performance for neuronal responses on a held-out test set: one focusing on predicting in-domain natural stimuli and another on out-of-distribution (OOD) stimuli to assess model generalization. As part of the NeurIPS 2023 competition track, we received more than 160 model submissions from 22 teams. Several new architectures for predictive models were proposed, and the winning teams improved the previous state-of-the-art model by 50%. Access to the dataset as well as the benchmarking infrastructure will remain online at www.sensorium-competition.net.

With the exponential growth of data and evolving use cases, petabyte-scale OLAP data platforms are increasingly adopting a model that decouples compute from storage. This shift, evident in organizations like Uber and Meta, introduces operational challenges including massive, read-heavy I/O traffic with potential throttling, as well as skewed and fragmented data access patterns. Addressing these challenges, this paper introduces the Alluxio local (edge) cache, a highly effective architectural optimization tailored for such environments. This embeddable cache, optimized for petabyte-scale data analytics, leverages local SSD resources to alleviate network I/O and API call pressures, significantly improving data transfer efficiency. Integrated with OLAP systems like Presto and storage services like HDFS, the Alluxio local cache has demonstrated its effectiveness in handling large-scale, enterprise-grade workloads over three years of deployment at Uber and Meta. We share insights and operational experiences in implementing these optimizations, providing valuable perspectives on managing modern, massive-scale OLAP workloads., Comment: Accepted to the USENIX Annual Technical Conference (USENIX ATC) 2024

Most existing multi-object tracking methods typically learn visual tracking features via maximizing dis-similarities of different instances and minimizing similarities of the same instance. While such a feature learning scheme achieves promising performance, learning discriminative features solely based on visual information is challenging especially in case of environmental interference such as occlusion, blur and domain variance. In this work, we argue that multi-modal language-driven features provide complementary information to classical visual features, thereby aiding in improving the robustness to such environmental interference. To this end, we propose a new multi-object tracking framework, named LG-MOT, that explicitly leverages language information at different levels of granularity (scene-and instance-level) and combines it with standard visual features to obtain discriminative representations. To develop LG-MOT, we annotate existing MOT datasets with scene-and instance-level language descriptions. We then encode both instance-and scene-level language information into high-dimensional embeddings, which are utilized to guide the visual features during training. At inference, our LG-MOT uses the standard visual features without relying on annotated language descriptions. Extensive experiments on three benchmarks, MOT17, DanceTrack and SportsMOT, reveal the merits of the proposed contributions leading to state-of-the-art performance. On the DanceTrack test set, our LG-MOT achieves an absolute gain of 2.2\% in terms of target object association (IDF1 score), compared to the baseline using only visual features. Further, our LG-MOT exhibits strong cross-domain generalizability. The dataset and code will be available at ~\url{https://github.com/WesLee88524/LG-MOT}.

Continuous normalizing flows (CNFs) learn an ordinary differential equation to transform prior samples into data. Flow matching (FM) has recently emerged as a simulation-free approach for training CNFs by regressing a velocity model towards the conditional velocity field. However, on constrained domains, the learned velocity model may lead to undesirable flows that result in highly unnatural samples, e.g., oversaturated images, due to both flow matching error and simulation error. To address this, we add a boundary constraint term to CNFs, which leads to reflected CNFs that keep trajectories within the constrained domains. We propose reflected flow matching (RFM) to train the velocity model in reflected CNFs by matching the conditional velocity fields in a simulation-free manner, similar to the vanilla FM. Moreover, the analytical form of conditional velocity fields in RFM avoids potentially biased approximations, making it superior to existing score-based generative models on constrained domains. We demonstrate that RFM achieves comparable or better results on standard image benchmarks and produces high-quality class-conditioned samples under high guidance weight., Comment: ICML 2024 camera-ready

Recently, unconventional antiferromagnets that enable the splitting of electronic spins have been theoretically proposed and experimentally realized, where the magnetic sublattices containing moments pointing at different directions are connected by a novel set of symmetries. Such spin splitting (SS) is substantial, $k$-dependent, and independent of the spin-orbit coupling strength, making these magnets promising materials for antiferromagnetic spintronics. Here, combined with angle-resolved photoemission spectroscopy (ARPES) and density functional theory (DFT) calculations, we perform a systematic study on CrSb, a metallic spin-split antiferromagnet candidate with $T_N$ = 703 K. Our data reveals the electronic structure of CrSb along both out-of-plane and in-plane momentum directions, which renders anisotropic $k$-dependent SS and agrees well with the calculational results. The magnitude of such SS reaches up to at least 0.8 eV at non-high-symmetry momentum points, which is significantly higher than the largest known SOC-induced SS. This compound expands the choice of materials in the field of antiferromagnetic spintronics and is likely to stimulate subsequent investigations of high-efficiency spintronic devices that are functional at room temperature., Comment: 14 pages, 4 figures

Cell image segmentation is usually implemented using fully supervised deep learning methods, which heavily rely on extensive annotated training data. Yet, due to the complexity of cell morphology and the requirement for specialized knowledge, pixel-level annotation of cell images has become a highly labor-intensive task. To address the above problems, we propose an active learning framework for cell segmentation using bounding box annotations, which greatly reduces the data annotation cost of cell segmentation algorithms. First, we generate a box-supervised learning method (denoted as YOLO-SAM) by combining the YOLOv8 detector with the Segment Anything Model (SAM), which effectively reduces the complexity of data annotation. Furthermore, it is integrated into an active learning framework that employs the MC DropBlock method to train the segmentation model with fewer box-annotated samples. Extensive experiments demonstrate that our model saves more than ninety percent of data annotation time compared to mask-supervised deep learning methods.

This paper discusses the results of the third edition of the Monocular Depth Estimation Challenge (MDEC). The challenge focuses on zero-shot generalization to the challenging SYNS-Patches dataset, featuring complex scenes in natural and indoor settings. As with the previous edition, methods can use any form of supervision, i.e. supervised or self-supervised. The challenge received a total of 19 submissions outperforming the baseline on the test set: 10 among them submitted a report describing their approach, highlighting a diffused use of foundational models such as Depth Anything at the core of their method. The challenge winners drastically improved 3D F-Score performance, from 17.51% to 23.72%., Comment: To appear in CVPRW2024

This paper reviews the NTIRE 2024 low light image enhancement challenge, highlighting the proposed solutions and results. The aim of this challenge is to discover an effective network design or solution capable of generating brighter, clearer, and visually appealing results when dealing with a variety of conditions, including ultra-high resolution (4K and beyond), non-uniform illumination, backlighting, extreme darkness, and night scenes. A notable total of 428 participants registered for the challenge, with 22 teams ultimately making valid submissions. This paper meticulously evaluates the state-of-the-art advancements in enhancing low-light images, reflecting the significant progress and creativity in this field., Comment: NTIRE 2024 Challenge Report

Matching cost aggregation plays a fundamental role in learning-based multi-view stereo networks. However, directly aggregating adjacent costs can lead to suboptimal results due to local geometric inconsistency. Related methods either seek selective aggregation or improve aggregated depth in the 2D space, both are unable to handle geometric inconsistency in the cost volume effectively. In this paper, we propose GoMVS to aggregate geometrically consistent costs, yielding better utilization of adjacent geometries. More specifically, we correspond and propagate adjacent costs to the reference pixel by leveraging the local geometric smoothness in conjunction with surface normals. We achieve this by the geometric consistent propagation (GCP) module. It computes the correspondence from the adjacent depth hypothesis space to the reference depth space using surface normals, then uses the correspondence to propagate adjacent costs to the reference geometry, followed by a convolution for aggregation. Our method achieves new state-of-the-art performance on DTU, Tanks & Temple, and ETH3D datasets. Notably, our method ranks 1st on the Tanks & Temple Advanced benchmark., Comment: CVPR 2024. Project page: https://wuuu3511.github.io/gomvs/ Code: https://github.com/Wuuu3511/GoMVS

Mixed solid compounds are employed in a vast array of applications so an accurate determination of their chemical compositions is of crucial importance. All current characterization methods require specially-treated samples so the availability of a more practical method with similar accuracy should alleviate the quantification process. In this work, we show how the doping concentration $\delta$ (or isotope concentration) of a mixed solid compound in powdered form, where both parent compounds are diamagnetic, can be obtained from the measurement of the mass magnetization. We exploit the additive nature of the molar magnetic susceptibility $\chi_{Mol}$ and molar mass to construct two equations with the same two unknowns in the $\chi_{Mol}$ vs. $\delta$ space to simultaneously solve $\chi_{Mol}$ and $\delta$ of a mixed solid. Eight examples are provided to show the wide applicability of this method: NH$_{4(1-\delta)}$D$_{4\delta}$Br (where D = $^2$H), NH$_4$I$_{1-\delta}$Br$_\delta$, (NH$_4$H$_2$)$_{1-\delta}$(ND$_4$D$_2$)$_\delta$PO$_4$, C$_{48}$H$_{22+6\delta}$Br$_{6(1-\delta)}$O$_{32}$Zr$_6$, [creatine]$_{1-\delta}$[$_D$-glucose]$_\delta$, [$_L$-glutamic acid]$_{1-\delta}$[$_L$-leucine]$_\delta$, [terephthalic acid]$_{1-\delta}$[trimesic acid]$_\delta$ and [p-terphenyl]$_{1-\delta}$[triphenylphosphine]$_\delta$. Experimental errors of ~1.2% were obtained for $\delta$ from average sample masses of 16.6 mg in powdered form rendering the presented approach an attractive choice for characterizing the ratios of mixed solids., Comment: Main article: 19 pages, 12 Figures; Supplementary Information: 7 pages, 9 Tables and 4 Figures

Reinforcement Learning from Human Feedback (RLHF) is the prevailing approach to ensure Large Language Models (LLMs) align with human values. However, existing RLHF methods require a high computational cost, one main reason being that RLHF assigns both the generation and alignment tasks to the LLM simultaneously. In this paper, we introduce Proxy-RLHF, which decouples the generation and alignment processes of LLMs, achieving alignment with human values at a much lower computational cost. We start with a novel Markov Decision Process (MDP) designed for the alignment process and employ Reinforcement Learning (RL) to train a streamlined proxy model that oversees the token generation of the LLM, without altering the LLM itself. Experiments show that our method achieves a comparable level of alignment with only 1\% of the training parameters of other methods.

The evolving landscape of online multiplayer gaming presents unique challenges in assessing the causal impacts of game features. Traditional A/B testing methodologies fall short due to complex player interactions, leading to violations of fundamental assumptions like the Stable Unit Treatment Value Assumption (SUTVA). Unlike traditional social networks with stable and long-term connections, networks in online games are often dynamic and short-lived. Players are temporarily teamed up for the duration of a game, forming transient networks that dissolve once the game ends. This fleeting nature of interactions presents a new challenge compared with running experiments in a stable social network. This study introduces a novel framework for treatment effect estimation in online gaming environments, considering the dynamic and ephemeral network interference that occurs among players. We propose an innovative estimator tailored for scenarios where a completely randomized experimental design is implemented without explicit knowledge of network structures. Notably, our method facilitates post-hoc interference adjustment on experimental data, significantly reducing the complexities and costs associated with intricate experimental designs and randomization strategies. The proposed framework stands out for its ability to accommodate varying levels of interference, thereby yielding more accurate and robust estimations. Through comprehensive simulations set against a variety of interference scenarios, along with empirical validation using real-world data from a mobile gaming environment, we demonstrate the efficacy of our approach. This study represents a pioneering effort in exploring causal inference in user-randomized experiments impacted by dynamic network effects.

Pairwise matching cost aggregation is a crucial step for modern learning-based Multi-view Stereo (MVS). Prior works adopt an early aggregation scheme, which adds up pairwise costs into an intermediate cost. However, we analyze that this process can degrade informative pairwise matchings, thereby blocking the depth network from fully utilizing the original geometric matching cues. To address this challenge, we present a late aggregation approach that allows for aggregating pairwise costs throughout the network feed-forward process, achieving accurate estimations with only minor changes of the plain CasMVSNet. Instead of building an intermediate cost by weighted sum, late aggregation preserves all pairwise costs along a distinct view channel. This enables the succeeding depth network to fully utilize the crucial geometric cues without loss of cost fidelity. Grounded in the new aggregation scheme, we propose further techniques addressing view order dependence inside the preserved cost, handling flexible testing views, and improving the depth filtering process. Despite its technical simplicity, our method improves significantly upon the baseline cascade-based approach, achieving comparable results with state-of-the-art methods with favorable computation overhead., Comment: Code and models are available at https://github.com/Wuuu3511/LAMVSNET

Recent studies have made remarkable progress in histopathology classification. Based on current successes, contemporary works proposed to further upgrade the model towards a more generalizable and robust direction through incrementally learning from the sequentially delivered domains. Unlike previous parameter isolation based approaches that usually demand massive computation resources during model updating, we present a memory-efficient prompt tuning framework to cultivate model generalization potential in economical memory cost. For each incoming domain, we reuse the existing parameters of the initial classification model and attach lightweight trainable prompts into it for customized tuning. Considering the domain heterogeneity, we perform decoupled prompt tuning, where we adopt a domain-specific prompt for each domain to independently investigate its distinctive characteristics, and one domain-invariant prompt shared across all domains to continually explore the common content embedding throughout time. All domain-specific prompts will be appended to the prompt bank and isolated from further changes to prevent forgetting the distinctive features of early-seen domains. While the domain-invariant prompt will be passed on and iteratively evolve by style-augmented prompt refining to improve model generalization capability over time. In specific, we construct a graph with existing prompts and build a style-augmented graph attention network to guide the domain-invariant prompt exploring the overlapped latent embedding among all delivered domains for more domain generic representations. We have extensively evaluated our framework with two histopathology tasks, i.e., breast cancer metastasis classification and epithelium-stroma tissue classification, where our approach yielded superior performance and memory efficiency over the competing methods., Comment: Accepted by AAAI 2024

FPGAs are increasingly prevalent in cloud deployments, serving as Smart NICs or network-attached accelerators. Despite their potential, developing distributed FPGA-accelerated applications remains cumbersome due to the lack of appropriate infrastructure and communication abstractions. To facilitate the development of distributed applications with FPGAs, in this paper we propose ACCL+, an open-source versatile FPGA-based collective communication library. Portable across different platforms and supporting UDP, TCP, as well as RDMA, ACCL+ empowers FPGA applications to initiate direct FPGA-to-FPGA collective communication. Additionally, it can serve as a collective offload engine for CPU applications, freeing the CPU from networking tasks. It is user-extensible, allowing new collectives to be implemented and deployed without having to re-synthesize the FPGA circuit. We evaluated ACCL+ on an FPGA cluster with 100 Gb/s networking, comparing its performance against software MPI over RDMA. The results demonstrate ACCL+'s significant advantages for FPGA-based distributed applications and highly competitive performance for CPU applications. We showcase ACCL+'s dual role with two use cases: seamlessly integrating as a collective offload engine to distribute CPU-based vector-matrix multiplication, and serving as a crucial and efficient component in designing fully FPGA-based distributed deep-learning recommendation inference.

We construct the complex multi-matrix model with W-representation and calculate the correlators. We establish the correspondence between the connected correlators and length-2n q-colored Dyck walks in Fredkin spin chain and discuss the entanglement entropy. Moreover, we analyze the free energy of this multi-matrix model. For the leading coefficient of the free energy, it relates to the connected correlators in large N limit.

Time-optimal control for high-order chain-of-integrators systems with full state constraints and arbitrarily given terminal states remains a challenging problem in the optimal control theory domain, yet to be resolved. To enhance further comprehension of the problem, this paper establishes a novel notation system and theoretical framework, providing the switching manifold for high-order problems in the form of switching laws. Through deriving properties of switching laws regarding signs and dimension, this paper proposes a definite condition for time-optimal control. Guided by the developed theory, a trajectory planning method named the manifold-intercept method (MIM) is developed. The proposed MIM can plan time-optimal jerk-limited trajectories with full state constraints, and can also plan near-optimal non-chattering higher-order trajectories with negligible extra motion time compared to optimal profiles. Numerical results indicate that the proposed MIM outperforms all baselines in computational time, computational accuracy, and trajectory quality by a large gap.

We study two-loop corrections to the scattering amplitude of four massive leptons in quantum electrodynamics. These amplitudes involve previously unknown elliptic Feynman integrals, which we compute analytically using the differential equation method. In doing so, we uncover the details of the elliptic geometry underlying this scattering amplitude and show how to exploit its properties to obtain compact, easy-to-evaluate series expansions that describe the scattering of four massive leptons in QED in the kinematical regions relevant for Bhabha and M{\o}ller scattering processes., Comment: 9 pages, 3 figures

The Joint Detection and Embedding (JDE) framework has achieved remarkable progress for multiple object tracking. Existing methods often employ extracted embeddings to re-establish associations between new detections and previously disrupted tracks. However, the reliability of embeddings diminishes when the region of the occluded object frequently contains adjacent objects or clutters, especially in scenarios with severe occlusion. To alleviate this problem, we propose a novel multiple object tracking method based on visual embedding consistency, mainly including: 1) Occlusion Prediction Module (OPM) and 2) Occlusion-Aware Association Module (OAAM). The OPM predicts occlusion information for each true detection, facilitating the selection of valid samples for consistency learning of the track's visual embedding. The OAAM leverages occlusion cues and visual embeddings to generate two separate embeddings for each track, guaranteeing consistency in both unoccluded and occluded detections. By integrating these two modules, our method is capable of addressing track interruptions caused by occlusion in online tracking scenarios. Extensive experimental results demonstrate that our approach achieves promising performance levels in both unoccluded and occluded tracking scenarios.

High Dynamic Range (HDR) images can be recovered from several Low Dynamic Range (LDR) images by existing Deep Neural Networks (DNNs) techniques. Despite the remarkable progress, DNN-based methods still generate ghosting artifacts when LDR images have saturation and large motion, which hinders potential applications in real-world scenarios. To address this challenge, we formulate the HDR deghosting problem as an image generation that leverages LDR features as the diffusion model's condition, consisting of the feature condition generator and the noise predictor. Feature condition generator employs attention and Domain Feature Alignment (DFA) layer to transform the intermediate features to avoid ghosting artifacts. With the learned features as conditions, the noise predictor leverages a stochastic iterative denoising process for diffusion models to generate an HDR image by steering the sampling process. Furthermore, to mitigate semantic confusion caused by the saturation problem of LDR images, we design a sliding window noise estimator to sample smooth noise in a patch-based manner. In addition, an image space loss is proposed to avoid the color distortion of the estimated HDR results. We empirically evaluate our model on benchmark datasets for HDR imaging. The results demonstrate that our approach achieves state-of-the-art performances and well generalization to real-world images., Comment: accepted by IEEE TCSVT