Search

The bottleneck associated with the key-value(KV) cache presents a significant challenge during the inference processes of large language models. While depth pruning accelerates inference, it requires extensive recovery training, which can take up to two weeks. On the other hand, width pruning retains much of the performance but offers slight speed gains. To tackle these challenges, we propose KVPruner to improve model efficiency while maintaining performance. Our method uses global perplexity-based analysis to determine the importance ratio for each block and provides multiple strategies to prune non-essential KV channels within blocks. Compared to the original model, KVPruner reduces runtime memory usage by 50% and boosts throughput by over 35%. Additionally, our method requires only two hours of LoRA fine-tuning on small datasets to recover most of the performance.

Determinant quantum Monte Carlo (DQMC), formulated in complex-fermion representation, has played a key role in studying strongly-correlated fermion systems. However, its applicability is limited due to the requirement of particle-number conservation after Hubbard-Stratonovich transformation. In going beyond the conventional DQMC, one encouraging development occurred when Majorana fermions were introduced for QMC [1,2]. But in previous Majorana-based QMC, Boltzmann weight is determined often with a sign ambiguity. Here we successfully resolved this ambiguity by deriving a closed-form Pfaffian formula for the weight, enabling efficient calculation of the weight with its sign in polynomial time. We call it ''Pfaffian quantum Monte Carlo'' (PfQMC), which can be applied to generic interacting fermion models. We have successfully employed PfQMC to explore how robust Majorana edge modes in Kitaev chain are against strong interactions. By offering greater flexibility, PfQMC can potentially enhance existing sign-mitigating and approximation methods and help address challenging issues such as the ground-state properties of the doped Hubbard model., Comment: 18 pages (including supplemental materials), 5 figures

The application of Large Language Models (LLMs) in healthcare is expanding rapidly, with one potential use case being the translation of formal medical reports into patient-legible equivalents. Currently, LLM outputs often need to be edited and evaluated by a human to ensure both factual accuracy and comprehensibility, and this is true for the above use case. We aim to minimize this step by proposing an agentic workflow with the Reflexion framework, which uses iterative self-reflection to correct outputs from an LLM. This pipeline was tested and compared to zero-shot prompting on 16 randomized radiology reports. In our multi-agent approach, reports had an accuracy rate of 94.94% when looking at verification of ICD-10 codes, compared to zero-shot prompted reports, which had an accuracy rate of 68.23%. Additionally, 81.25% of the final reflected reports required no corrections for accuracy or readability, while only 25% of zero-shot prompted reports met these criteria without needing modifications. These results indicate that our approach presents a feasible method for communicating clinical findings to patients in a quick, efficient and coherent manner whilst also retaining medical accuracy. The codebase is available for viewing at http://github.com/malavikhasudarshan/Multi-Agent-Patient-Letter-Generation., Comment: 12 pages, 7 figures

Binary verification plays a pivotal role in software security, yet building a verification service that is both open and trustworthy poses a formidable challenge. In this paper, we introduce a novel binary verification service, AGORA, scrupulously designed to overcome the challenge. At the heart of this approach lies a strategic insight: certain tasks can be delegated to untrusted entities, while the corresponding validators are securely housed within the trusted computing base (TCB). AGORA can validate untrusted assertions generated for versatile policies. Through a novel blockchain-based bounty task manager, it also utilizes crowdsourcing to remove trust in theorem provers. These synergistic techniques successfully ameliorate the TCB size burden associated with two procedures: binary analysis and theorem proving. The design of AGORA allows untrusted parties to participate in these complex processes. Moreover, based on running the optimized TCB within trusted execution environments and recording the verification process on a blockchain, the public can audit the correctness of verification results. By implementing verification workflows for software-based fault isolation policy and side-channel mitigation, our evaluation demonstrates the efficacy of AGORA.

We report new Lyman-$\alpha$ and 21-cm constraints on freeze-in dark matter (FIDM) which injects energy into the intergalactic medium either through annihilation or decay to photon(s) or electron-positron pair. With respect to Lyman-$\alpha$ we fix the baseline ionization history using low redshift data about astrophysical reionization, whereas for 21-cm signal we adopt the baseline values of 21-cm power spectrum through a standard modeling of star formation developed so far. Using the latest numerical tools, we show that (i) for sterile neutrino FIDM, current Lyman-$\alpha$ data and future sensitivity of SKA-low (1000 hrs) on the 21-cm power spectra excludes the FIDM mass up to $1.8\times 10^{-3}$ GeV at 95$\%$ CL and $5.46\times 10^{-4}$ GeV, respectively, and (ii) for millicharged FIDM, current Lyman-$\alpha$ data only excludes the millicharge down to $10^{-8}$ within the FIDM mass range of $10^{-3}-1$ GeV at 95$\%$ CL, suggesting that the surviving parameter space of millicharged FIDM is still intact., Comment: 17 pages, 5 figures; v2: revised version with 19 pp and 8 figs

Cooperative adaptive cruise control (CACC) has been recognized as a fundamental function of autonomous driving, in which platoon stability and energy efficiency are outstanding challenges that are difficult to accommodate in real-world operations. This paper studied the CACC of connected and autonomous vehicles (CAVs) based on the multi-agent reinforcement learning algorithm (MARL) to optimize platoon stability and energy efficiency simultaneously. The optimal use of communication bandwidth is the key to guaranteeing learning performance in real-world driving, and thus this paper proposes a communication-efficient MARL by incorporating the quantified stochastic gradient descent (QSGD) and a binary differential consensus (BDC) method into a fully-decentralized MARL framework. We benchmarked the performance of our proposed BDC-MARL algorithm against several several non-communicative andcommunicative MARL algorithms, e.g., IA2C, FPrint, and DIAL, through the evaluation of platoon stability, fuel economy, and driving comfort. Our results show that BDC-MARL achieved the highest energy savings, improving by up to 5.8%, with an average velocity of 15.26 m/s and an inter-vehicle spacing of 20.76 m. In addition, we conducted different information-sharing analyses to assess communication efficacy, along with sensitivity analyses and scalability tests with varying platoon sizes. The practical effectiveness of our approach is further demonstrated using real-world scenarios sourced from open-sourced OpenACC.

Convergence analysis of Markov chain Monte Carlo methods in high-dimensional statistical applications is increasingly recognized. In this paper, we develop general mixing time bounds for Metropolis-Hastings algorithms on discrete spaces by building upon and refining some recent theoretical advancements in Bayesian model selection problems. We establish sufficient conditions for a class of informed Metropolis-Hastings algorithms to attain relaxation times that are independent of the problem dimension. These conditions are grounded in high-dimensional statistical theory and allow for possibly multimodal posterior distributions. We obtain our results through two independent techniques: the multicommodity flow method and single-element drift condition analysis; we find that the latter yields a tighter mixing time bound. Our results and proof techniques are readily applicable to a broad spectrum of statistical problems with discrete parameter spaces.

Legal retrieval techniques play an important role in preserving the fairness and equality of the judicial system. As an annually well-known international competition, COLIEE aims to advance the development of state-of-the-art retrieval models for legal texts. This paper elaborates on the methodology employed by the TQM team in COLIEE2024.Specifically, we explored various lexical matching and semantic retrieval models, with a focus on enhancing the understanding of case relevance. Additionally, we endeavor to integrate various features using the learning-to-rank technique. Furthermore, fine heuristic pre-processing and post-processing methods have been proposed to mitigate irrelevant information. Consequently, our methodology achieved remarkable performance in COLIEE2024, securing first place in Task 1 and third place in Task 3. We anticipate that our proposed approach can contribute valuable insights to the advancement of legal retrieval technology., Comment: 16 pages

Schr\"{o}dinger bridge can be viewed as a continuous-time stochastic control problem where the goal is to find an optimally controlled diffusion process whose terminal distribution coincides with a pre-specified target distribution. We propose to generalize this problem by allowing the terminal distribution to differ from the target but penalizing the Kullback-Leibler divergence between the two distributions. We call this new control problem soft-constrained Schr\"{o}dinger bridge (SSB). The main contribution of this work is a theoretical derivation of the solution to SSB, which shows that the terminal distribution of the optimally controlled process is a geometric mixture of the target and some other distribution. This result is further extended to a time series setting. One application is the development of robust generative diffusion models. We propose a score matching-based algorithm for sampling from geometric mixtures and showcase its use via a numerical example for the MNIST data set., Comment: Made minor changes about the references. 38 pages, 7 figures. Accepted by AISTATS 2024

Out-of-distribution detection (OOD) is a crucial technique for deploying machine learning models in the real world to handle the unseen scenarios. In this paper, we first propose a simple yet effective Neural Activation Prior (NAP) for OOD detection. Our neural activation prior is based on a key observation that, for a channel before the global pooling layer of a fully trained neural network, the probability of a few neurons being activated with a large response by an in-distribution (ID) sample is significantly higher than that by an OOD sample. An intuitive explanation is that for a model fully trained on ID dataset, each channel would play a role in detecting a certain pattern in the ID dataset, and a few neurons can be activated with a large response when the pattern is detected in an input sample. Then, a new scoring function based on this prior is proposed to highlight the role of these strongly activated neurons in OOD detection. Our approach is plug-and-play and does not lead to any performance degradation on ID data classification and requires no extra training or statistics from training or external datasets. Notice that previous methods primarily rely on post-global-pooling features of the neural networks, while the within-channel distribution information we leverage would be discarded by the global pooling operator. Consequently, our method is orthogonal to existing approaches and can be effectively combined with them in various applications. Experimental results show that our method achieves the state-of-the-art performance on CIFAR benchmark and ImageNet dataset, which demonstrates the power of the proposed prior. Finally, we extend our method to Transformers and the experimental findings indicate that NAP can also significantly enhance the performance of OOD detection on Transformers, thereby demonstrating the broad applicability of this prior knowledge.

Due to its innovative weighting method, the high-order target ENO (TENO) system has shown tremendous potential for complicated flow prediction. For unstructured meshes, we have developed non-oscillatory central target ENO family schemes in this study. In comparison with standard WENO schemes, the compact directional stencils significantly enhance the possibility that one of them will be in smooth area. The scheme is inherently limited in size in order to simplify the directional stencils in implementation. Following an effective scale separation method, we employ the ENO-like stencil selection technique, which selects candidate reconstructions from large central stencils in smooth areas to enforce the ENO property and from small directional stencils at discontinuities to restore high-order accuracy. With the validation of a number of test cases, a detailed comparison of CWENO, TENO, CTENO and CTENOZ schemes was carried out to assess their performance based on accuracy, robustness, parallel scalability and run time. Our findings demonstrate high-order precision, lower numerical dissipation, and superior sharp shock-capturing performance of proposed CTENO and CTENOZ schemes.

Connected and automated vehicles (CAVs) have emerged as a potential solution to the future challenges of developing safe, efficient, and eco-friendly transportation systems. However, CAV control presents significant challenges, given the complexity of interconnectivity and coordination required among the vehicles. To address this, multi-agent reinforcement learning (MARL), with its notable advancements in addressing complex problems in autonomous driving, robotics, and human-vehicle interaction, has emerged as a promising tool for enhancing the capabilities of CAVs. However, there is a notable absence of current reviews on the state-of-the-art MARL algorithms in the context of CAVs. Therefore, this paper delivers a comprehensive review of the application of MARL techniques within the field of CAV control. The paper begins by introducing MARL, followed by a detailed explanation of its unique advantages in addressing complex mobility and traffic scenarios that involve multiple agents. It then presents a comprehensive survey of MARL applications on the extent of control dimensions for CAVs, covering critical and typical scenarios such as platooning control, lane-changing, and unsignalized intersections. In addition, the paper provides a comprehensive review of the prominent simulation platforms used to create reliable environments for training in MARL. Lastly, the paper examines the current challenges associated with deploying MARL within CAV control and outlines potential solutions that can effectively overcome these issues. Through this review, the study highlights the tremendous potential of MARL to enhance the performance and collaboration of CAV control in terms of safety, travel efficiency, and economy.

Practical extreme ultraviolet (EUV) sources yield the desired 13.5 nm radiation but also generate debris, significantly limiting the lifespan of the collector mirror in lithography. In this study, we explore the role of buffer gas in transporting debris particles within a EUV source vessel using direct numerical simulations (DNS). Our study involves a 2m $\times$ 1m $\times$ 1m rectangular cavity with an injecting jet flow subjected to sideward outlet. Debris particles are introduced into the cavity with specified initial velocities, simulating a spherical radiating pattern with particle diameters ranging from 0.1 $\mu$m to 1 $\mu$m. Varying the inflow velocity (from $1$m/s to $50$m/s) of the buffer gas reveals a morphological transition in the flow field. At low inflow velocities, the flow remains steady, whereas higher inflow velocities induce the formation of clustered corner rolls. Upon reaching sufficiently high inflow velocities, the jet flow can penetrate the entire cavity, impacting the endwall. Interestingly, the resulting recirculation flow leads to the spontaneous formation of spiraling outflow. The distinct flow structures at various inflow velocities lead to distinct patterns of particle transport. For low-speed gas, it is efficient in expelling all particles smaller than 0.4 $\mu$m, while for high-speed gas, those fine particles accumulate near the endwall and are challenging to be extracted. Our findings highlight the significance of controlling flow conditions for effective debris particle transport and clearance in diverse applications especially in EUV source vessels.

Clustering of time series is a well-studied problem, with applications ranging from quantitative, personalized models of metabolism obtained from metabolite concentrations to state discrimination in quantum information theory. We consider a variant, where given a set of trajectories and a number of parts, we jointly partition the set of trajectories and learn linear dynamical system (LDS) models for each part, so as to minimize the maximum error across all the models. We present globally convergent methods and EM heuristics, accompanied by promising computational results.

Fairness holds a pivotal role in the realm of machine learning, particularly when it comes to addressing groups categorised by sensitive attributes, e.g., gender, race. Prevailing algorithms in fair learning predominantly hinge on accessibility or estimations of these sensitive attributes, at least in the training process. We design a single group-blind projection map that aligns the feature distributions of both groups in the source data, achieving (demographic) group parity, without requiring values of the protected attribute for individual samples in the computation of the map, as well as its use. Instead, our approach utilises the feature distributions of the privileged and unprivileged groups in a boarder population and the essential assumption that the source data are unbiased representation of the population. We present numerical results on synthetic data and real data.

Commonly used datasets for evaluating video codecs are all very high quality and not representative of video typically used in video conferencing scenarios. We present the Video Conferencing Dataset (VCD) for evaluating video codecs for real-time communication, the first such dataset focused on video conferencing. VCD includes a wide variety of camera qualities and spatial and temporal information. It includes both desktop and mobile scenarios and two types of video background processing. We report the compression efficiency of H.264, H.265, H.266, and AV1 in low-delay settings on VCD and compare it with the non-video conferencing datasets UVC, MLC-JVC, and HEVC. The results show the source quality and the scenarios have a significant effect on the compression efficiency of all the codecs. VCD enables the evaluation and tuning of codecs for this important scenario. The VCD is publicly available as an open-source dataset at https://github.com/microsoft/VCD.

The growing adoption of hybrid electric vehicles (HEVs) presents a transformative opportunity for revolutionizing transportation energy systems. The shift towards electrifying transportation aims to curb environmental concerns related to fossil fuel consumption. This necessitates efficient energy management systems (EMS) to optimize energy efficiency. The evolution of EMS from HEVs to connected hybrid electric vehicles (CHEVs) represent a pivotal shift. For HEVs, EMS now confronts the intricate energy cooperation requirements of CHEVs, necessitating advanced algorithms for route optimization, charging coordination, and load distribution. Challenges persist in both domains, including optimal energy utilization for HEVs, and cooperative eco-driving control (CED) for CHEVs across diverse vehicle types. Reinforcement learning (RL) stands out as a promising tool for addressing these challenges. Specifically, within the realm of CHEVs, the application of multi-agent reinforcement learning (MARL) emerges as a powerful approach for effectively tackling the intricacies of CED control. Despite extensive research, few reviews span from individual vehicles to multi-vehicle scenarios. This review bridges the gap, highlighting challenges, advancements, and potential contributions of RL-based solutions for future sustainable transportation systems.

Testing of function safety and Safety Of The Intended Functionality (SOTIF) is important for autonomous vehicles (AVs). It is hard to test the AV's hazard response in the real world because it would involve hazards to passengers and other road users. This paper studied on virtual testing of AV on the CARLA platform and proposed a Humanoid Scenario Generation (HSG) scheme to investigate the impacts of hazardous behaviors on AV collision rates. The HSG scheme breakthrough the current limitation on the rarity and reproducibility of real scenes. By accurately capturing five prominent human driver behaviors that directly contribute to vehicle collisions in the real world, the methodology significantly enhances the realism and diversity of the simulation, as evidenced by collision rate statistics across various traffic scenarios. Thus, the modular framework allows for customization, and its seamless integration within the CARLA platform ensures compatibility with existing tools. Ultimately, the comparison results demonstrate that all vehicles that exhibited hazardous behaviors followed the predefined random speed distribution and the effectiveness of the HSG was validated by the distinct characteristics displayed by these behaviors.

We present the Parareal-CG algorithm for time-dependent differential equations in this work. The algorithm is a parallel in time iteration algorithm utilizes Chebyshev-Gauss spectral collocation method for fine propagator F and backward Euler method for coarse propagator G. As far as we know, this is the first time that the spectral method used as the F propagator of the parareal algorithm. By constructing the stable function of the Chebyshev-Gauss spectral collocation method for the symmetric positive definite (SPD) problem, we find out that the Parareal-CG algorithm and the Parareal-TR algorithm, whose F propagator is chosen to be a trapezoidal ruler, converge similarly, i.e., the Parareal-CG algorithm converge as fast as Parareal-Euler algorithm with sufficient Chebyhsev-Gauss points in every coarse grid. Numerical examples including ordinary differential equations and time-dependent partial differential equations are given to illustrate the high efficiency and accuracy of the proposed algorithm.

Informed importance tempering (IIT) is an easy-to-implement MCMC algorithm that can be seen as an extension of the familiar Metropolis-Hastings algorithm with the special feature that informed proposals are always accepted, and which was shown in Zhou and Smith (2022) to converge much more quickly in some common circumstances. This work develops a new, comprehensive guide to the use of IIT in many situations. First, we propose two IIT schemes that run faster than existing informed MCMC methods on discrete spaces by not requiring the posterior evaluation of all neighboring states. Second, we integrate IIT with other MCMC techniques, including simulated tempering, pseudo-marginal and multiple-try methods (on general state spaces), which have been conventionally implemented as Metropolis-Hastings schemes and can suffer from low acceptance rates. The use of IIT allows us to always accept proposals and brings about new opportunities for optimizing the sampler which are not possible under the Metropolis-Hastings framework. Numerical examples illustrating our findings are provided for each proposed algorithm, and a general theory on the complexity of IIT methods is developed.