Search

The increasing popularity of deep learning models has created new opportunities for developing AI-based recommender systems. Designing recommender systems using deep neural networks requires careful architecture design, and further optimization demands extensive co-design efforts on jointly optimizing model architecture and hardware. Design automation, such as Automated Machine Learning (AutoML), is necessary to fully exploit the potential of recommender model design, including model choices and model-hardware co-design strategies. We introduce a novel paradigm that utilizes weight sharing to explore abundant solution spaces. Our paradigm creates a large supernet to search for optimal architectures and co-design strategies to address the challenges of data multi-modality and heterogeneity in the recommendation domain. From a model perspective, the supernet includes a variety of operators, dense connectivity, and dimension search options. From a co-design perspective, it encompasses versatile Processing-In-Memory (PIM) configurations to produce hardware-efficient models. Our solution space's scale, heterogeneity, and complexity pose several challenges, which we address by proposing various techniques for training and evaluating the supernet. Our crafted models show promising results on three Click-Through Rates (CTR) prediction benchmarks, outperforming both manually designed and AutoML-crafted models with state-of-the-art performance when focusing solely on architecture search. From a co-design perspective, we achieve 2x FLOPs efficiency, 1.8x energy efficiency, and 1.5x performance improvements in recommender models., Comment: Accepted in ACM Transactions on Recommender Systems. arXiv admin note: substantial text overlap with arXiv:2207.07187

Large language models (LLMs) have achieved substantial progress in processing long contexts but still struggle with long-context reasoning. Existing approaches typically involve fine-tuning LLMs with synthetic data, which depends on annotations from human experts or advanced models like GPT-4, thus restricting further advancements. To address this issue, we investigate the potential for LLMs to self-improve in long-context reasoning and propose \ours, an approach specifically designed for this purpose. This approach is straightforward: we sample multiple outputs for each question, score them with Minimum Bayes Risk, and then apply supervised fine-tuning or preference optimization based on these outputs. Extensive experiments on several leading LLMs demonstrate the effectiveness of \ours, with an absolute improvement of $4.2$ points for Llama-3.1-8B-Instruct. Furthermore, \ours achieves superior performance compared to prior approaches that depend on data produced by human experts or advanced models. We anticipate that this work will open new avenues for self-improvement techniques in long-context scenarios, which are essential for the continual advancement of LLMs., Comment: Project Page: https://github.com/SihengLi99/SEALONG

This paper presents a monitoring framework that infers the level of autonomous vehicle (AV) collision risk based on its object detector's performance using only monocular camera images. Essentially, the framework takes two sets of predictions produced by different algorithms and associates their inconsistencies with the collision risk via fuzzy inference. The first set of predictions is obtained through retrieving safety-critical 2.5D objects from a depth map, and the second set comes from the AV's 3D object detector. We experimentally validate that, based on Intersection-over-Union (IoU) and a depth discrepancy measure, the inconsistencies between the two sets of predictions strongly correlate to the safety-related error of the 3D object detector against ground truths. This correlation allows us to construct a fuzzy inference system and map the inconsistency measures to an existing collision risk indicator. In particular, we apply various knowledge- and data-driven techniques and find using particle swarm optimization that learns general fuzzy rules gives the best mapping result. Lastly, we validate our monitor's capability to produce relevant risk estimates with the large-scale nuScenes dataset and show it can safeguard an AV in closed-loop simulations., Comment: 7 pages (IEEE double column format), 5 figures, 3 tables, submitted to ICRA 2025

Existing research on non-line-of-sight (NLoS) ultraviolet (UV) channel modeling mainly focuses on scenarios where the signal propagation process is not affected by any obstacle and the radiation intensity (RI) of the light source is uniformly distributed. To eliminate these restrictions, we propose a single-collision model for the NLoS UV channel incorporating a cuboid-shaped obstacle, where the RI of the UV light source is modeled as the Lambertian distribution. For easy interpretation, we categorize the intersection circumstances between the receiver field-of-view and the obstacle into six cases and provide derivations of the weighting factor for each case. To investigate the accuracy of the proposed model, we compare it with the associated Monte Carlo photon tracing model via simulations and experiments. Results verify the correctness of the proposed model. This work reveals that obstacle avoidance is not always beneficial for NLoS UV communications and provides guidelines for relevant system design., Comment: Submitted to IEEE International Conference on Communications (ICC) 2025

Multimodal foundation models, such as Gemini and ChatGPT, have revolutionized human-machine interactions by seamlessly integrating various forms of data. Developing a universal spoken language model that comprehends a wide range of natural language instructions is critical for bridging communication gaps and facilitating more intuitive interactions. However, the absence of a comprehensive evaluation benchmark poses a significant challenge. We present Dynamic-SUPERB Phase-2, an open and evolving benchmark for the comprehensive evaluation of instruction-based universal speech models. Building upon the first generation, this second version incorporates 125 new tasks contributed collaboratively by the global research community, expanding the benchmark to a total of 180 tasks, making it the largest benchmark for speech and audio evaluation. While the first generation of Dynamic-SUPERB was limited to classification tasks, Dynamic-SUPERB Phase-2 broadens its evaluation capabilities by introducing a wide array of novel and diverse tasks, including regression and sequence generation, across speech, music, and environmental audio. Evaluation results indicate that none of the models performed well universally. SALMONN-13B excelled in English ASR, while WavLLM demonstrated high accuracy in emotion recognition, but current models still require further innovations to handle a broader range of tasks. We will soon open-source all task data and the evaluation pipeline.

Neutral atom array has emerged as a promising platform for quantum computation owing to its high-fidelity two-qubit gate, arbitrary connectivity and overwhelming scalability. Nevertheless, fault-tolerant quantum computing on the neutral atom platform requires consideration of the types of errors that neutral atoms are prone to. One typical and major error is leakage error from Rydberg state when implementing multi-qubit gate. Such leakage error is harmful by propagating multiple pauli errors in quantum circuit. Researchers have proposed erasure conversion protocol, which utilizes fast leakage detection to convert leakage error to benign erasure error. This method has a favorable error distance d, but is limited to certain atom species. Here, we propose a new method to deal with such leakage error in measurement-based quantum computation (MBQC), to which we refer as "Leakage Tracking". We remove the demand for mid-circuit leakage detection but infer the probabilities and locations of pauli errors through gate sequence and final leakage detection. We show that this method has an error distance de = d and reaches a high threshold 1.7% per CZ gate for pure leakage error and perfect final leakage detection. In presence of atom loss and other pauli errors, we show the advantage in error distance over erasure conversion when the ratio of leakage error is close to one., Comment: 11 pages, 5 figures

Implicit Neural Representations (INRs) have emerged as a paradigm in knowledge representation, offering exceptional flexibility and performance across a diverse range of applications. INRs leverage multilayer perceptrons (MLPs) to model data as continuous implicit functions, providing critical advantages such as resolution independence, memory efficiency, and generalisation beyond discretised data structures. Their ability to solve complex inverse problems makes them particularly effective for tasks including audio reconstruction, image representation, 3D object reconstruction, and high-dimensional data synthesis. This survey provides a comprehensive review of state-of-the-art INR methods, introducing a clear taxonomy that categorises them into four key areas: activation functions, position encoding, combined strategies, and network structure optimisation. We rigorously analyse their critical properties, such as full differentiability, smoothness, compactness, and adaptability to varying resolutions while also examining their strengths and limitations in addressing locality biases and capturing fine details. Our experimental comparison offers new insights into the trade-offs between different approaches, showcasing the capabilities and challenges of the latest INR techniques across various tasks. In addition to identifying areas where current methods excel, we highlight key limitations and potential avenues for improvement, such as developing more expressive activation functions, enhancing positional encoding mechanisms, and improving scalability for complex, high-dimensional data. This survey serves as a roadmap for researchers, offering practical guidance for future exploration in the field of INRs. We aim to foster new methodologies by outlining promising research directions for INRs and applications.

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

The Water Cherenkov Detector Array (WCDA) is one of the components of Large High Altitude Air Shower Observatory (LHAASO) and can monitor any sources over two-thirds of the sky for up to 7 hours per day with >98\% duty cycle. In this work, we report the detection of two outbursts of the Fanaroff-Riley I radio galaxy NGC 1275 that were detected by LHAASO-WCDA between November 2022 and January 2023 with statistical significance of 5.2~$\sigma$ and 8.3~$\sigma$. The observed spectral energy distribution in the range from 500 GeV to 3 TeV is fitted by a power-law with a best-fit spectral index of $\alpha=-3.37\pm0.52$ and $-3.35\pm0.29$, respectively. The outburst flux above 0.5~TeV was ($4.55\pm 4.21)\times~10^{-11}~\rm cm^{-2}~s^{-1}$ and ($3.45\pm 1.78)\times~10^{-11}~\rm cm^{-2}~s^{-1}$, corresponding to 60\%, 45\% of Crab Nebula flux. Variation analysis reveals the variability time-scale of days at the TeV energy band. A simple test by one-zone synchrotron self-Compton model reproduces the data in the gamma-ray band well., Comment: 11 pages, 8 figures, 3 tables

The emergence of Mixture of Experts (MoE) LLMs has significantly advanced the development of language models. Compared to traditional LLMs, MoE LLMs outperform traditional LLMs by achieving higher performance with considerably fewer activated parameters. Despite this efficiency, their enormous parameter size still leads to high deployment costs. In this paper, we introduce a two-stage compression method tailored for MoE to reduce the model size and decrease the computational cost. First, in the inter-expert pruning stage, we analyze the importance of each layer and propose the Layer-wise Genetic Search and Block-wise KT-Reception Field with the non-uniform pruning ratio to prune the individual expert. Second, in the intra-expert decomposition stage, we apply the low-rank decomposition to further compress the parameters within the remaining experts. Extensive experiments on Qwen1.5-MoE-A2.7B, DeepSeek-V2-Lite, and Mixtral-8$\times$7B demonstrate that our proposed methods can both reduce the model size and enhance inference efficiency while maintaining performance in various zero-shot tasks. The code will be available at \url{https://github.com/xiaochengsky/MoEI-2.git}

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)

Multimodal Question Answering (MMQA) is crucial as it enables comprehensive understanding and accurate responses by integrating insights from diverse data representations such as tables, charts, and text. Most existing researches in MMQA only focus on two modalities such as image-text QA, table-text QA and chart-text QA, and there remains a notable scarcity in studies that investigate the joint analysis of text, tables, and charts. In this paper, we present C$\text{T}^2$C-QA, a pioneering Chinese reasoning-based QA dataset that includes an extensive collection of text, tables, and charts, meticulously compiled from 200 selectively sourced webpages. Our dataset simulates real webpages and serves as a great test for the capability of the model to analyze and reason with multimodal data, because the answer to a question could appear in various modalities, or even potentially not exist at all. Additionally, we present AED (\textbf{A}llocating, \textbf{E}xpert and \textbf{D}esicion), a multi-agent system implemented through collaborative deployment, information interaction, and collective decision-making among different agents. Specifically, the Assignment Agent is in charge of selecting and activating expert agents, including those proficient in text, tables, and charts. The Decision Agent bears the responsibility of delivering the final verdict, drawing upon the analytical insights provided by these expert agents. We execute a comprehensive analysis, comparing AED with various state-of-the-art models in MMQA, including GPT-4. The experimental outcomes demonstrate that current methodologies, including GPT-4, are yet to meet the benchmarks set by our dataset., Comment: 10 pages, 6 figures

GPT-4o is an autoregressive omni model that accepts as input any combination of text, audio, image, and video, and generates any combination of text, audio, and image outputs. It's trained end-to-end across text, vision, and audio, meaning all inputs and outputs are processed by the same neural network. GPT-4o can respond to audio inputs in as little as 232 milliseconds, with an average of 320 milliseconds, which is similar to human response time in conversation. It matches GPT-4 Turbo performance on text in English and code, with significant improvement on text in non-English languages, while also being much faster and 50\% cheaper in the API. GPT-4o is especially better at vision and audio understanding compared to existing models. In line with our commitment to building AI safely and consistent with our voluntary commitments to the White House, we are sharing the GPT-4o System Card, which includes our Preparedness Framework evaluations. In this System Card, we provide a detailed look at GPT-4o's capabilities, limitations, and safety evaluations across multiple categories, focusing on speech-to-speech while also evaluating text and image capabilities, and measures we've implemented to ensure the model is safe and aligned. We also include third-party assessments on dangerous capabilities, as well as discussion of potential societal impacts of GPT-4o's text and vision capabilities.

In this work, we propose a unified representation for Super-Resolution (SR) and Image Compression, termed **Factorized Fields**, motivated by the shared principles between these two tasks. Both SISR and Image Compression require recovering and preserving fine image details--whether by enhancing resolution or reconstructing compressed data. Unlike previous methods that mainly focus on network architecture, our proposed approach utilizes a basis-coefficient decomposition to explicitly capture multi-scale visual features and structural components in images, addressing the core challenges of both tasks. We first derive our SR model, which includes a Coefficient Backbone and Basis Swin Transformer for generalizable Factorized Fields. Then, to further unify these two tasks, we leverage the strong information-recovery capabilities of the trained SR modules as priors in the compression pipeline, improving both compression efficiency and detail reconstruction. Additionally, we introduce a merged-basis compression branch that consolidates shared structures, further optimizing the compression process. Extensive experiments show that our unified representation delivers state-of-the-art performance, achieving an average relative improvement of 204.4% in PSNR over the baseline in Super-Resolution (SR) and 9.35% BD-rate reduction in Image Compression compared to the previous SOTA., Comment: Project page: https://jayisaking.github.io/FIPER/

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

For superconducting quantum processors, stable high-fidelity two-qubit operations depend on precise flux control of the tunable coupler. However, the pulse distortion poses a significant challenge to the control precision. Current calibration methods, which often rely on microwave crosstalk or additional readout resonators for coupler excitation and readout, tend to be cumbersome and inefficient, especially when couplers only have flux control. Here, we introduce and experimentally validate a novel pulse calibration scheme that exploits the strong coupling between qubits and couplers, eliminating the need for extra coupler readout and excitation. Our method directly measures the short-time and long-time step responses of the coupler flux pulse transient, enabling us to apply predistortion to subsequent signals using fast Fourier transformation and deconvolution. This approach not only simplifies the calibration process but also significantly improves the precision and stability of the flux control. We demonstrate the efficacy of our method through the implementation of diabatic CZ and iSWAP gates with fidelities of $99.61\pm0.04\%$ and $99.82\pm0.02\%$, respectively, as well as a series of diabatic CPhase gates with high fidelities characterized by cross-entropy benchmarking. The consistency and robustness of our technique are further validated by the reduction in pulse distortion and phase error observed across multilayer CZ gates. These results underscore the potential of our calibration and predistortion method to enhance the performance of two-qubit gates in superconducting quantum processors., Comment: 16 pages, 11 figures, 7 tables

Approximate Nearest Neighbor Search (ANNS) is now widely used in various applications, ranging from information retrieval, question answering, and recommendation, to search for similar high-dimensional vectors. As the amount of vector data grows continuously, it becomes important to support updates to vector index, the enabling technique that allows for efficient and accurate ANNS on vectors. Because of the curse of high dimensionality, it is often costly to identify the right neighbors of a single new vector, a necessary process for index update. To amortize update costs, existing systems maintain a secondary index to accumulate updates, which are merged by the main index by global rebuilding the entire index periodically. However, this approach has high fluctuations of search latency and accuracy, not even to mention that it requires substantial resources and is extremely time-consuming for rebuilds. We introduce SPFresh, a system that supports in-place vector updates. At the heart of SPFresh is LIRE, a lightweight incremental rebalancing protocol to split vector partitions and reassign vectors in the nearby partitions to adapt to data distribution shift. LIRE achieves low-overhead vector updates by only reassigning vectors at the boundary between partitions, where in a high-quality vector index the amount of such vectors are deemed small. With LIRE, SPFresh provides superior query latency and accuracy to solutions based on global rebuild, with only 1% of DRAM and less than 10% cores needed at the peak compared to the state-of-the-art, in a billion scale vector index with 1% of daily vector update rate., Comment: SOSP 23

In 1927, during the fifth Solvay Conference, Einstein and Bohr described a double-slit interferometer with a "movable slit" that can detect the momentum recoil of one photon. Here, we report a faithful realization of the Einstein-Bohr interferometer using a single atom in an optical tweezer, cooled to the motional ground state in three dimensions. The single atom has an intrinsic momentum uncertainty comparable to a single photon, which serves as a movable slit obeying the minimum Heisenberg uncertainty principle. The atom's momentum wavefunction is dynamically tunable by the tweezer laser power, which enables observation of an interferometric visibility reduction at a shallower trap, demonstrating the quantum nature of this interferometer. We further identify classical noise due to atom heating and precession, illustrating a quantum-to-classical transition., Comment: 18 pages, 4 figures

Echocardiography is the most widely used cardiac imaging modality, capturing ultrasound video data to assess cardiac structure and function. Artificial intelligence (AI) in echocardiography has the potential to streamline manual tasks and improve reproducibility and precision. However, most echocardiography AI models are single-view, single-task systems that do not synthesize complementary information from multiple views captured during a full exam, and thus lead to limited performance and scope of applications. To address this problem, we introduce EchoPrime, a multi-view, view-informed, video-based vision-language foundation model trained on over 12 million video-report pairs. EchoPrime uses contrastive learning to train a unified embedding model for all standard views in a comprehensive echocardiogram study with representation of both rare and common diseases and diagnoses. EchoPrime then utilizes view-classification and a view-informed anatomic attention model to weight video-specific interpretations that accurately maps the relationship between echocardiographic views and anatomical structures. With retrieval-augmented interpretation, EchoPrime integrates information from all echocardiogram videos in a comprehensive study and performs holistic comprehensive clinical echocardiography interpretation. In datasets from two independent healthcare systems, EchoPrime achieves state-of-the art performance on 23 diverse benchmarks of cardiac form and function, surpassing the performance of both task-specific approaches and prior foundation models. Following rigorous clinical evaluation, EchoPrime can assist physicians in the automated preliminary assessment of comprehensive echocardiography., Comment: 30 pages, 3 tables, 3 figures

In this paper, we consider global strong solutions and uniform-in-time vanishing damping limit for the inviscid Oldroyd-B model in R^d, where d=2 and 3. The well-recognized problem of the global existence of smooth solutions for the 2D inviscid Oldroyd-B model without smallness assumptions is open due to the complex structure of Q. Therefore improving the smallness assumptions, especially in lower regularity class, is the core question in the area of fluid models. On the other hand, long-time behaviors of solutions including temporal decay and uniform-in-time damping stability are also of deep significance. These problems have been widely studied, however, the existing results are not regularity critical and the (uniform) vanishing damping limit has not been discussed. The goal of this work is to dig deeper in this direction. In this work we first establish the local well-posedness in the sense of Hadamard with critical regularity. Then, by virtue of the sharp commutator estimate for Calderon-Zygmund operator, we establish the global existence of solutions for d=2 with damping in the low regularity class, which to our best knowledge, is novel in the literature. Furthermore, in both 2D and 3D cases, we prove the global existence of the solutions to the inviscid Oldroyd-B model independent of the damping parameters. In addition, we obtain the optimal temporal decay rates and time integrability by improving the existing Fourier splitting method and developing a novel decomposition strategy. One of the major contributions of the presenting paper is to prove the uniform-in-time vanishing damping limit for the inviscid Oldroyd-B model and discover the correlation between sharp vanishing damping rate and the temporal decay rate. Finally, we will support our findings by providing numerical evidence regarding the vanishing damping limit in the periodic domain T^d., Comment: 91 pages, 14 figures

Action recognition has witnessed the development of a growing number of novel algorithms and datasets in the past decade. However, the majority of public benchmarks were constructed around activities of daily living and annotated at a rather coarse-grained level, which lacks diversity in domain-specific datasets, especially for rarely seen domains. In this paper, we introduced Human Stone Toolmaking Action Grammar (HSTAG), a meticulously annotated video dataset showcasing previously undocumented stone toolmaking behaviors, which can be used for investigating the applications of advanced artificial intelligence techniques in understanding a rapid succession of complex interactions between two hand-held objects. HSTAG consists of 18,739 video clips that record 4.5 hours of experts' activities in stone toolmaking. Its unique features include (i) brief action durations and frequent transitions, mirroring the rapid changes inherent in many motor behaviors; (ii) multiple angles of view and switches among multiple tools, increasing intra-class variability; (iii) unbalanced class distributions and high similarity among different action sequences, adding difficulty in capturing distinct patterns for each action. Several mainstream action recognition models are used to conduct experimental analysis, which showcases the challenges and uniqueness of HSTAG https://nyu.databrary.org/volume/1697., Comment: 8 pages, 4 figures, accepted by the 11th IEEE International Conference on Data Science and Advanced Analytics (DSAA)

This paper examines the phenomenon of probabilistic robustness overestimation in TRADES, a prominent adversarial training method. Our study reveals that TRADES sometimes yields disproportionately high PGD validation accuracy compared to the AutoAttack testing accuracy in the multiclass classification task. This discrepancy highlights a significant overestimation of robustness for these instances, potentially linked to gradient masking. We further analyze the parameters contributing to unstable models that lead to overestimation. Our findings indicate that smaller batch sizes, lower beta values (which control the weight of the robust loss term in TRADES), larger learning rates, and higher class complexity (e.g., CIFAR-100 versus CIFAR-10) are associated with an increased likelihood of robustness overestimation. By examining metrics such as the First-Order Stationary Condition (FOSC), inner-maximization, and gradient information, we identify the underlying cause of this phenomenon as gradient masking and provide insights into it. Furthermore, our experiments show that certain unstable training instances may return to a state without robust overestimation, inspiring our attempts at a solution. In addition to adjusting parameter settings to reduce instability or retraining when overestimation occurs, we recommend incorporating Gaussian noise in inputs when the FOSC score exceed the threshold. This method aims to mitigate robustness overestimation of TRADES and other similar methods at its source, ensuring more reliable representation of adversarial robustness during evaluation.

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.

Cosmic muon imaging technology is increasingly being applied in various fields. However, simulating cosmic muons typically requires the rapid generation of a large number of muons and tracking their complex trajectories through intricate structures. This process is highly computationally demanding and consumes significant CPU time. To address these challenges, we introduce DeepMuon, an innovative deep learning model designed to efficiently and accurately generate cosmic muon distributions. In our approach, we employ the inverse Box-Cox transformation to reduce the kurtosis of the muon energy distribution, making it more statistically manageable for the model to learn. Additionally, we utilize the Sliced Wasserstein Distance (SWD) as a loss function to ensure precise simulation of the high-dimensional distributions of cosmic muons. We also demonstrate that DeepMuon can accurately learn muon distribution patterns from a limited set of data, enabling it to simulate real-world cosmic muon distributions as captured by detectors. Compared to traditional tools like CRY, DeepMuon significantly increases the speed of muon generation at sea level. Furthermore, we have developed a pipeline using DeepMuon that directly simulates muon distributions in underwater environments, dramatically accelerating simulations for underwater muon radiography and tomography. For more details on our open-source project, please visit https://github.com/wangab0/deepmuon.

The first very long baseline interferometry (VLBI) detections at 870$\mu$m wavelength (345$\,$GHz frequency) are reported, achieving the highest diffraction-limited angular resolution yet obtained from the surface of the Earth, and the highest-frequency example of the VLBI technique to date. These include strong detections for multiple sources observed on inter-continental baselines between telescopes in Chile, Hawaii, and Spain, obtained during observations in October 2018. The longest-baseline detections approach 11$\,$G$\lambda$ corresponding to an angular resolution, or fringe spacing, of 19$\mu$as. The Allan deviation of the visibility phase at 870$\mu$m is comparable to that at 1.3$\,$mm on the relevant integration time scales between 2 and 100$\,$s. The detections confirm that the sensitivity and signal chain stability of stations in the Event Horizon Telescope (EHT) array are suitable for VLBI observations at 870$\mu$m. Operation at this short wavelength, combined with anticipated enhancements of the EHT, will lead to a unique high angular resolution instrument for black hole studies, capable of resolving the event horizons of supermassive black holes in both space and time., Comment: Corresponding author: S. Doeleman

We report the detection of an extended very-high-energy (VHE) gamma-ray source coincident with the locations of middle-aged (62.4~\rm kyr) pulsar PSR J0248+6021, by using the LHAASO-WCDA data of live 796 days and LHAASO-KM2A data of live 1216 days. A significant excess of \gray induced showers is observed both by WCDA in energy bands of 1-25~\rm TeV and KM2A in energy bands of $>$ 25~\rm TeV with 7.3 $\sigma$ and 13.5 $\sigma$, respectively. The best-fit position derived through WCDA data is R.A. = 42.06$^\circ \pm$ 0.12$^\circ$ and Dec. = 60.24$^\circ \pm $ 0.13$^\circ$ with an extension of 0.69$^\circ\pm$0.15$^\circ$ and that of the KM2A data is R.A.= 42.29$^\circ \pm $ 0.13$^\circ$ and Dec. = 60.38$^\circ \pm$ 0.07$^\circ$ with an extension of 0.37$^\circ\pm$0.07$^\circ$. No clear extended multiwavelength counterpart of this LHAASO source has been found from the radio band to the GeV band. The most plausible explanation of the VHE \gray emission is the inverse Compton process of highly relativistic electrons and positrons injected by the pulsar. These electrons/positrons are hypothesized to be either confined within the pulsar wind nebula or to have already escaped into the interstellar medium, forming a pulsar halo., Comment: 12 pages, 10 figures, Accepted by Sci. China-Phys. Mech. Astron

Traditional vertical take-off and landing (VTOL) aircraft can not achieve optimal efficiency for various payload weights and has limited mobility due to its under-actuation. With the thrust-vectoring mechanism, the proposed modular team UAV is fully actuated at certain attitudes. However, the attainable force space (AFS) differs according to the team configuration, which makes the controller design difficult. We propose an approximation to the AFS and a full-pose tracking controller with an attitude planner and a force projection, which guarantees the control force is feasible. The proposed approach can be applied to UAVs having multiple thrust-vectoring effectors with homogeneous agents. The simulation and experiment demonstrate a tilting motion during hovering for a 4-agent team.

The in vivo propulsion and monitoring of nanoparticles (NPs) have received tremendous achievements in the past decade. Developing functional NPs that can be efficiently manipulated inside the human body with a non-invasive tracking modality is critical to clinical translation. This study synthesized a photo-magnetically powered nanoparticle (PMN) with a Fe3O4 core and gold spiky surface. The Au-nanotips ensure PMNs have a strong light absorption in the second near-infrared (NIR) window and produce outstanding photoacoustic signals. The Bio-transmission electron microscopy and simulation results prove that the assembly of PMNs under a magnetic field further enhances the photothermal conversion in cells, contributing to the reduction of ambient viscosity. Photoacoustic imaging (PAI) realized real-time monitoring of PMN movements and revealed that laser plus magnetic coupling couldimprove intratumoral distribution and retention. The proposed methods exhibit excellent potential for the clinical research of cancer nanotherapies.

Self-consistency-based approaches, which involve repeatedly sampling multiple outputs and selecting the most consistent one as the final response, prove to be remarkably effective in improving the factual accuracy of large language models. Nonetheless, existing methods usually have strict constraints on the task format, largely limiting their applicability. In this paper, we present Integrative Decoding (ID), to unlock the potential of self-consistency in open-ended generation tasks. ID operates by constructing a set of inputs, each prepended with a previously sampled response, and then processes them concurrently, with the next token being selected by aggregating of all their corresponding predictions at each decoding step. In essence, this simple approach implicitly incorporates self-consistency in the decoding objective. Extensive evaluation shows that ID consistently enhances factuality over a wide range of language models, with substantial improvements on the TruthfulQA (+11.2%), Biographies (+15.4%) and LongFact (+8.5%) benchmarks. The performance gains amplify progressively as the number of sampled responses increases, indicating the potential of ID to scale up with repeated sampling.

Deaf and Hard-of-Hearing (DHH) learners face unique challenges in video-based learning due to the complex interplay between visual and auditory information in videos. Traditional approaches to making video content accessible primarily focus on captioning, but these solutions often neglect the cognitive demands of processing both visual and textual information simultaneously. This paper introduces a set of \textit{Motion} design guidelines, aimed at mitigating these cognitive challenges and improving video learning experiences for DHH learners. Through a two-phase research, we identified five key challenges, including misaligned content and visual overload. We proposed five design principles accordingly. User study with 16 DHH participants showed that improving visual-audio relevance and guiding visual attention significantly enhances the learning experience by reducing physical demand, alleviating temporal pressure, and improving learning satisfaction. Our findings highlight the potential of Motion design to transform educational content for DHH learners, and we discuss implications for inclusive video learning tools.

Honesty is a fundamental principle for aligning large language models (LLMs) with human values, requiring these models to recognize what they know and don't know and be able to faithfully express their knowledge. Despite promising, current LLMs still exhibit significant dishonest behaviors, such as confidently presenting wrong answers or failing to express what they know. In addition, research on the honesty of LLMs also faces challenges, including varying definitions of honesty, difficulties in distinguishing between known and unknown knowledge, and a lack of comprehensive understanding of related research. To address these issues, we provide a survey on the honesty of LLMs, covering its clarification, evaluation approaches, and strategies for improvement. Moreover, we offer insights for future research, aiming to inspire further exploration in this important area., Comment: Project Page: https://github.com/SihengLi99/LLM-Honesty-Survey

We propose an innovative scheme to efficiently prepare strong mechanical squeezing through utilizing the synergistic mechanism of two-tone driving and parametric pumping in an optomechanical system. By reasonable choosing the system parameters, the proposal highlights the following prominent advantages: the squeezing effect of the cavity field induced by the optical parametric amplifier can be transferred to the mechanical oscillator, which has been squeezed by the two-tone driving, and the degree of squeezing of the mechanical oscillator will surpass that obtained by any single mechanism; the joint mechanism can enhance the degree of squeezing significantly and break the 3 dB mechanical squeezing limit, which is particularly evident in range where the red/blue-detuned ratio is sub-optimal; the mechanical squeezing achieved through this distinctive joint mechanism exhibits notable robustness against both thermal noise and decay of mechanical oscillator. Our project offers a versatile and efficient approach for generating strong mechanical squeezing across a wide range of conditions.

We study a center-of-mass-conserving Brownian complex Sachdev-Ye-Kitaev model with long-range (power-law) interactions characterized by $1/r^\eta$. The kinetic constraint and long-range interactions conspire to yield rich hydrodynamics associated with the conserved charge, which we reveal by computing the Schwinger-Keldysh effective action. Our result shows that charge transport in this system can be subdiffusive, diffusive, or superdiffusive, with the dynamical exponent controlled by $\eta$. We further employ a doubled Hilbert space methodology to derive an effective action for the out-of-time-order correlator (OTOC), from which we obtain the phase diagram delineating regimes where the lightcone is linear or logarithmic. Our results provide a concrete example of a quantum many-body system with kinetic constraint and long-range interactions in which the emergent hydrodynamic modes and OTOC can be computed analytically., Comment: 4.5 + 10 pages, 4 figures

Fluorite-type $\mathrm{HfO_2}$-based ferroelectric (FE) oxides have rekindled interest in FE memories due to their compatibility with silicon processing and potential for high-density integration. The polarization characteristics of FE devices are governed by the dynamics of metastable domain structure evolution. Insightful design of FE devices for encoding and storage necessitates a comprehensive understanding of the internal structural evolution. Here, we demonstrate the evolution of domain structures through a transient polar orthorhombic (O)-$Pmn2_1$-like configuration via $in$-$situ$ biasing on $\mathrm{TiN/Hf_{0.5}Zr_{0.5}O_2/TiN}$ capacitors within spherical aberration-corrected transmission electron microscope, combined with theoretical calculations. Furthermore, it is directly evidenced that the non-FE O-$Pbca$ transforms into the FE O-$Pca2_1$ phase under electric field, with the polar axis of the FE-phase aligning towards the bias direction through ferroelastic transformation, thereby enhancing FE polarization. As cycling progresses further, however, the polar axis collapses, leading to FE degradation. These novel insights into the intricate structural evolution path under electrical field cycling facilitate optimization and design strategies for $\mathrm{HfO_2}$-based FE memory devices., Comment: 16 pages, 4 figures for the main text

Speech Emotion Recognition (SER) is a crucial component in developing general-purpose AI agents capable of natural human-computer interaction. However, building robust multilingual SER systems remains challenging due to the scarcity of labeled data in languages other than English and Chinese. In this paper, we propose an approach to enhance SER performance in low SER resource languages by leveraging data from high-resource languages. Specifically, we employ expressive Speech-to-Speech translation (S2ST) combined with a novel bootstrapping data selection pipeline to generate labeled data in the target language. Extensive experiments demonstrate that our method is both effective and generalizable across different upstream models and languages. Our results suggest that this approach can facilitate the development of more scalable and robust multilingual SER systems., Comment: 5 pages, 2 figures, Submitted to ICASSP 2025

Robot navigation is increasingly crucial across applications like delivery services and warehouse management. The integration of Reinforcement Learning (RL) with classical planning has given rise to meta-planners that combine the adaptability of RL with the explainable decision-making of classical planners. However, the exploration capabilities of RL-based meta-planners during training are often constrained by the capabilities of the underlying classical planners. This constraint can result in limited exploration, thereby leading to sampling skew issues. To address these issues, our paper introduces a novel framework, DIGIMON, which begins with behavior-guided diagnosis for exploration bottlenecks within the meta-planner and follows up with a mitigation strategy that conducts up-sampling from diagnosed bottleneck data. Our evaluation shows 13.5%+ improvement in navigation performance, greater robustness in out-of-distribution environments, and a 4x boost in training efficiency. DIGIMON is designed as a versatile, plug-and-play solution, allowing seamless integration into various RL-based meta-planners.

This paper aims to explain how a deep neural network (DNN) gradually extracts new knowledge and forgets noisy features through layers in forward propagation. Up to now, although the definition of knowledge encoded by the DNN has not reached a consensus, Previous studies have derived a series of mathematical evidence to take interactions as symbolic primitive inference patterns encoded by a DNN. We extend the definition of interactions and, for the first time, extract interactions encoded by intermediate layers. We quantify and track the newly emerged interactions and the forgotten interactions in each layer during the forward propagation, which shed new light on the learning behavior of DNNs. The layer-wise change of interactions also reveals the change of the generalization capacity and instability of feature representations of a DNN.

This paper introduces CompressedMediQ, a novel hybrid quantum-classical machine learning pipeline specifically developed to address the computational challenges associated with high-dimensional multi-class neuroimaging data analysis. Standard neuroimaging datasets, such as large-scale MRI data from the Alzheimer's Disease Neuroimaging Initiative (ADNI) and Neuroimaging in Frontotemporal Dementia (NIFD), present significant hurdles due to their vast size and complexity. CompressedMediQ integrates classical high-performance computing (HPC) nodes for advanced MRI pre-processing and Convolutional Neural Network (CNN)-PCA-based feature extraction and reduction, addressing the limited-qubit availability for quantum data encoding in the NISQ (Noisy Intermediate-Scale Quantum) era. This is followed by Quantum Support Vector Machine (QSVM) classification. By utilizing quantum kernel methods, the pipeline optimizes feature mapping and classification, enhancing data separability and outperforming traditional neuroimaging analysis techniques. Experimental results highlight the pipeline's superior accuracy in dementia staging, validating the practical use of quantum machine learning in clinical diagnostics. Despite the limitations of NISQ devices, this proof-of-concept demonstrates the transformative potential of quantum-enhanced learning, paving the way for scalable and precise diagnostic tools in healthcare and signal processing.

As Autonomous driving systems (ADS) have transformed our daily life, safety of ADS is of growing significance. While various testing approaches have emerged to enhance the ADS reliability, a crucial gap remains in understanding the accidents causes. Such post-accident analysis is paramount and beneficial for enhancing ADS safety and reliability. Existing cyber-physical system (CPS) root cause analysis techniques are mainly designed for drones and cannot handle the unique challenges introduced by more complex physical environments and deep learning models deployed in ADS. In this paper, we address the gap by offering a formal definition of ADS root cause analysis problem and introducing ROCAS, a novel ADS root cause analysis framework featuring cyber-physical co-mutation. Our technique uniquely leverages both physical and cyber mutation that can precisely identify the accident-trigger entity and pinpoint the misconfiguration of the target ADS responsible for an accident. We further design a differential analysis to identify the responsible module to reduce search space for the misconfiguration. We study 12 categories of ADS accidents and demonstrate the effectiveness and efficiency of ROCAS in narrowing down search space and pinpointing the misconfiguration. We also show detailed case studies on how the identified misconfiguration helps understand rationale behind accidents., Comment: Accepted at ASE 2024

Automated speaking assessment in conversation tests (ASAC) aims to evaluate the overall speaking proficiency of an L2 (second-language) speaker in a setting where an interlocutor interacts with one or more candidates. Although prior ASAC approaches have shown promising performance on their respective datasets, there is still a dearth of research specifically focused on incorporating the coherence of the logical flow within a conversation into the grading model. To address this critical challenge, we propose a hierarchical graph model that aptly incorporates both broad inter-response interactions (e.g., discourse relations) and nuanced semantic information (e.g., semantic words and speaker intents), which is subsequently fused with contextual information for the final prediction. Extensive experimental results on the NICT-JLE benchmark dataset suggest that our proposed modeling approach can yield considerable improvements in prediction accuracy with respect to various assessment metrics, as compared to some strong baselines. This also sheds light on the importance of investigating coherence-related facets of spoken responses in ASAC., Comment: Accepted by IEEE SLT 2024

End-to-end (E2E) automatic speech recognition (ASR) models have become standard practice for various commercial applications. However, in real-world scenarios, the long-tailed nature of word distribution often leads E2E ASR models to perform well on common words but fall short in recognizing uncommon ones. Recently, the notion of a contextual adapter (CA) was proposed to infuse external knowledge represented by a context word list into E2E ASR models. Although CA can improve recognition performance on rare words, two crucial data imbalance problems remain. First, when using low-frequency words as context words during training, since these words rarely occur in the utterance, CA becomes prone to overfit on attending to the token due to higher-frequency words not being present in the context list. Second, the long-tailed distribution within the context list itself still causes the model to perform poorly on low-frequency context words. In light of this, we explore in-depth the impact of altering the context list to have words with different frequency distributions on model performance, and meanwhile extend CA with a simple yet effective context-balanced learning objective. A series of experiments conducted on the AISHELL-1 benchmark dataset suggests that using all vocabulary words from the training corpus as the context list and pairing them with our balanced objective yields the best performance, demonstrating a significant reduction in character error rate (CER) by up to 1.21% and a more pronounced 9.44% reduction in the error rate of zero-shot words., Comment: Accepted by SLT 2024

Understanding the mechanisms behind neural network optimization is crucial for improving network design and performance. While various optimization techniques have been developed, a comprehensive understanding of the underlying principles that govern these techniques remains elusive. Specifically, the role of symmetry breaking, a fundamental concept in physics, has not been fully explored in neural network optimization. This gap in knowledge limits our ability to design networks that are both efficient and effective. Here, we propose the symmetry breaking hypothesis to elucidate the significance of symmetry breaking in enhancing neural network optimization. We demonstrate that a simple input expansion can significantly improve network performance across various tasks, and we show that this improvement can be attributed to the underlying symmetry breaking mechanism. We further develop a metric to quantify the degree of symmetry breaking in neural networks, providing a practical approach to evaluate and guide network design. Our findings confirm that symmetry breaking is a fundamental principle that underpins various optimization techniques, including dropout, batch normalization, and equivariance. By quantifying the degree of symmetry breaking, our work offers a practical technique for performance enhancement and a metric to guide network design without the need for complete datasets and extensive training processes., Comment: 29 pages, 8 figures

With the success of Diffusion Models for image generation, the technologies also have revolutionized the aesthetic Quick Response (QR) code generation. Despite significant improvements in visual attractiveness for the beautified codes, their scannabilities are usually sacrificed and thus hinder their practical uses in real-world scenarios. To address this issue, we propose a novel Diffusion-based QR Code generator (DiffQRCoder) to effectively craft both scannable and visually pleasing QR codes. The proposed approach introduces Scanning-Robust Perceptual Guidance (SRPG), a new diffusion guidance for Diffusion Models to guarantee the generated aesthetic codes to obey the ground-truth QR codes while maintaining their attractiveness during the denoising process. Additionally, we present another post-processing technique, Scanning Robust Manifold Projected Gradient Descent (SR-MPGD), to further enhance their scanning robustness through iterative latent space optimization. With extensive experiments, the results demonstrate that our approach not only outperforms other compared methods in Scanning Success Rate (SSR) with better or comparable CLIP aesthetic score (CLIP-aes.) but also significantly improves the SSR of the ControlNet-only approach from 60% to 99%. The subjective evaluation indicates that our approach achieves promising visual attractiveness to users as well. Finally, even with different scanning angles and the most rigorous error tolerance settings, our approach robustly achieves over 95% SSR, demonstrating its capability for real-world applications.

Two-dimensional (2D) magnets have attracted significant attentions in recent years due to their importance in the research on both fundamental physics and spintronic applications. Here, we report the discovery of a new ternary compound FePd2Te2. It features a layered quasi-2D crystal structure with one-dimensional Fe zigzag chains extending along the b-axis in the cleavage plane. Single crystals of FePd2Te2 with centimeter-size could be grown. Density functional theory calculations, mechanical exfoliation and atomic force microscopy on these crystals reveal that they are 2D materialsthat can be thinned down to 5 nm. Magnetic characterization shows that FePd2Te2 is an easy-plane ferromagnet with Tc 183 K and strong in-plane uniaxial magnetic anisotropy. Magnetoresistance and anomalous Hall effect demonstrate that ferromagnetism could maintain in FePd2Te2 flakes with large coercivity. A crystal twinning effect is observed by scanning tunneling microscopy which makes the Fe chains right-angle bent in the cleavage plane and creates an intriguing spin texture. Our results show that FePd2Te2 is a correlated anisotropic 2D magnets that may attract multidisciplinary research interests.

This is an index of the contributions by the Giant Radio Array for Neutrino Detection (GRAND) Collaboration to the 10th International Workshop on Acoustic and Radio EeV Neutrino Detection Activities (ARENA 2024, University of Chicago, June 11-14, 2024). The contributions include an overview of GRAND in its present and future incarnations, methods of radio-detection that are being developed for them, and ongoing joint work between the GRAND and BEACON experiments., Comment: Note: To access the list of contributions, please follow the "HTML" link that can be found on the arXiv page

For decades, assessing income has served as the tried-and-true method for creating financial aid packages--scholarships, grants, and loans--for the nation's college students. Each year, students and families living with low and moderate incomes submit income documentation to colleges, states, and the federal government in hopes of qualifying for financial aid to defray college costs. But what if a student's wealth status revealed more about their ability to cover college costs than their income did? Income is an important measure, but it provides only a snapshot of the full resources available to students and their families. Wealth, too, impacts a student's access and ability to pay for college as well as their chances of graduating and turning career aspirations into a well-paying job and fulfilling life. Examining income alone masks deep and persistent inequities in wealth by race or ethnicity. This gap, which is rooted in generations of discrimination and systemic barriers to opportunity, affects the college dreams of countless people, particularly those from historically marginalized communities. Limitations in available data mean researchers and policymakers know relatively little about how wealth impacts higher education, with far more attention focused on understanding and addressing the effects of income. This paper seeks to address that missing piece by exploring the impact of wealth on students' experiences in higher education and how these experiences, in turn, influence wealth-building.