Search

Ensuring correctness is crucial for code generation. Formal verification offers a definitive assurance of correctness, but demands substantial human effort in proof construction and hence raises a pressing need for automation. The primary obstacle lies in the severe lack of data - there is much less proof than code for LLMs to train upon. In this paper, we introduce SAFE, a novel framework that overcomes the lack of human-written proof to enable automated proof generation of Rust code. SAFE establishes a self-evolving cycle where data synthesis and fine-tuning collaborate to enhance the model capability, leveraging the definitive power of a symbolic verifier in telling correct proof from incorrect ones. SAFE also re-purposes the large number of synthesized incorrect proofs to train the self-debugging capability of the fine-tuned models, empowering them to fix incorrect proofs based on the verifier's feedback. SAFE demonstrates superior efficiency and precision compared to GPT-4o. Through tens of thousands of synthesized proofs and the self-debugging mechanism, we improve the capability of open-source models, initially unacquainted with formal verification, to automatically write proof for Rust code. This advancement leads to a significant improvement in performance, achieving a 70.50% accuracy rate in a benchmark crafted by human experts, a significant leap over GPT-4o's performance of 24.46%.

Large Multimodal Models (LMMs) have achieved significant breakthroughs in various vision-language and vision-centric tasks based on auto-regressive modeling. However, these models typically focus on either vision-centric tasks, such as visual grounding and region description, or vision-language tasks, like image caption and multi-scenario VQAs. None of the LMMs have yet comprehensively unified both types of tasks within a single model, as seen in Large Language Models in the natural language processing field. Furthermore, even with abundant multi-task instruction-following data, directly stacking these data for universal capabilities extension remains challenging. To address these issues, we introduce a novel multi-dimension curated and consolidated multimodal dataset, named CCMD-8M, which overcomes the data barriers of unifying vision-centric and vision-language tasks through multi-level data curation and multi-task consolidation. More importantly, we present Griffon-G, a general large multimodal model that addresses both vision-centric and vision-language tasks within a single end-to-end paradigm. Griffon-G resolves the training collapse issue encountered during the joint optimization of these tasks, achieving better training efficiency. Evaluations across multimodal benchmarks, general Visual Question Answering (VQA) tasks, scene text-centric VQA tasks, document-related VQA tasks, Referring Expression Comprehension, and object detection demonstrate that Griffon-G surpasses the advanced LMMs and achieves expert-level performance in complicated vision-centric tasks., Comment: This work has been submitted to the IEEE for possible publication. Codes and data will be later released at https://github.com/jefferyZhan/Griffon

Line graph transformation has been widely studied in graph theory, where each node in a line graph corresponds to an edge in the original graph. This has inspired a series of graph neural networks (GNNs) applied to transformed line graphs, which have proven effective in various graph representation learning tasks. However, there is limited theoretical study on how line graph transformation affects the expressivity of GNN models. In this study, we focus on two types of graphs known to be challenging to the Weisfeiler-Leman (WL) tests: Cai-F\"urer-Immerman (CFI) graphs and strongly regular graphs, and show that applying line graph transformation helps exclude these challenging graph properties, thus potentially assist WL tests in distinguishing these graphs. We empirically validate our findings by conducting a series of experiments that compare the accuracy and efficiency of graph isomorphism tests and GNNs on both line-transformed and original graphs across these graph structure types., Comment: 21 pages, code available at https://github.com/lukeyf/graphs-and-lines

We introduce Baichuan Alignment, a detailed analysis of the alignment techniques employed in the Baichuan series of models. This represents the industry's first comprehensive account of alignment methodologies, offering valuable insights for advancing AI research. We investigate the critical components that enhance model performance during the alignment process, including optimization methods, data strategies, capability enhancements, and evaluation processes. The process spans three key stages: Prompt Augmentation System (PAS), Supervised Fine-Tuning (SFT), and Preference Alignment. The problems encountered, the solutions applied, and the improvements made are thoroughly recorded. Through comparisons across well-established benchmarks, we highlight the technological advancements enabled by Baichuan Alignment. Baichuan-Instruct is an internal model, while Qwen2-Nova-72B and Llama3-PBM-Nova-70B are instruct versions of the Qwen2-72B and Llama-3-70B base models, optimized through Baichuan Alignment. Baichuan-Instruct demonstrates significant improvements in core capabilities, with user experience gains ranging from 17% to 28%, and performs exceptionally well on specialized benchmarks. In open-source benchmark evaluations, both Qwen2-Nova-72B and Llama3-PBM-Nova-70B consistently outperform their respective official instruct versions across nearly all datasets. This report aims to clarify the key technologies behind the alignment process, fostering a deeper understanding within the community. Llama3-PBM-Nova-70B model is available at https://huggingface.co/PKU-Baichuan-MLSystemLab/Llama3-PBM-Nova-70B.

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

This paper analyzes the working principle of four-switch cross-shaped reconfigurable intelligent surface (RIS) in detail and reveals the different types of RIS that can be designed based on this structure. Combined with the design examples using this structure in the currently published articles, this paper summarizes and organizes them, and also points out several RIS solutions that have not been designed using this structure. Finally, based on this four-switch cross-shaped structure, this paper proposes a novel RIS design example that can realize the function switching of 1-bit ultra-wideband (UWB) and 2-bit narrowband, and conducts simulation verification. The simulation results show that by optimizing the element structure and controlling the states of the four switches, the 1-bit ultra-wideband function can achieve a frequency band coverage of 10.5GHz-19.8GHz and a 2-bit phase quantization function around 18.12GHz. At the same time, it can realize 60{\deg} two-dimensional beam scanning function. We call this novel design "bit reconfigurable metasurface".

In the new perspective of spatial quantization, this article systematically studies the advantages of reconfigurable reflectarray (RRA) designed with closely spaced elements in terms of sidelobe level (SLL), scanning accuracy, scan loss and beam granularity, including theoretical analysis and simulation verification. This article sequentially studies RRAs with element periods of {\lambda}/2, {\lambda}/4 and {\lambda}/8. Both theoretical and simulation results show that under the condition of the same aperture size, with the number of spatial quantization bits increasing, 1bit RRA using closely spaced structure SLL will have a improvement of about 5dB. The scanning accuracy at 60{\deg} is improved from 54.52{\deg} at {\lambda}/2 to 57.97{\deg} at {\lambda}/8, while the scan loss is improved from 5.02dB at {\lambda}/2 to 2.85dB at {\lambda}/8. In terms of beam granularity, the beam granularity is increased by about 4 times for every 1bit of spatial quantization encryption in the RRA element period. The beam granularity at 0{\deg} of 1bit RRA with unit period of {\lambda}/2 is 0.166{\deg}, {\lambda}/4 is 0.033{\deg}, and {\lambda}/8 is 0.009{\deg}. This study has an important reference value for reconfigurable reflectarray design, communication system and radar design.

The cloud cover and meteorological parameters serve as fundamental criteria for the qualification of an astronomical observatory working in optical and infrared wavelengths. In this paper, we present a systematic assessment of key meteorological parameters at the Lenghu site. The datasets adopted in this study includes the meteorological parameters collected at the local weather stations at the site and in the Lenghu Town, the sky brightness at the local zenith acquired by the Sky Quality Meters and night sky all-sky images from a digital camera, the ERA5 reanalysis database and global climate monitoring data. From 2019 to 2023, the fractional observable time of photometric condition is 69.70%, 74.97%, 70.26%, 74.27% and 65.12%, respectively. The fractional observing time is inversely correlated with surface air temperature, relative humidity, precipitable water vapor, and dew temperature, demonstrating that the observing conditions are influenced by these meteorological parameters. Large-scale air-sea interactions affect the climate at Lenghu site, which in fact delivers a clue to understand the irregularity of 2023. Specifically, precipitable water vapor at Lenghu site is correlated to both the westerly wind index and the summer North Atlantic Oscillation index, the yearly average temperature of Lenghu site is observed to increase significantly during the occurrence of a strong El Ni\~no event and the relative humidity anomaly at Lenghu site is correlated to the Pacific Decadal Oscillation index. The decrease of fractional observing time in 2023 was due to the ongoing strong El Ni\~no event and relevant global climate change. We underscore the substantial role of global climate change in regulating astronomical observing conditions and the necessity for long-term continuous monitoring of the astronomical meteorological parameters at Lenghu site., Comment: accepted for publication in MNRAS

Attention is the cornerstone of modern Large Language Models (LLMs). Yet its quadratic complexity limits the efficiency and scalability of LLMs, especially for those with a long-context window. A promising approach addressing this limitation is to leverage the sparsity in attention. However, existing sparsity-based solutions predominantly rely on predefined patterns or heuristics to approximate sparsity. This practice falls short to fully capture the dynamic nature of attention sparsity in language-based tasks. This paper argues that attention sparsity should be learned rather than predefined. To this end, we design SeerAttention, a new Attention mechanism that augments the conventional attention with a learnable gate that adaptively selects significant blocks in an attention map and deems the rest blocks sparse. Such block-level sparsity effectively balances accuracy and speedup. To enable efficient learning of the gating network, we develop a customized FlashAttention implementation that extracts the block-level ground truth of attention map with minimum overhead. SeerAttention not only applies to post-training, but also excels in long-context fine-tuning. Our results show that at post-training stages, SeerAttention significantly outperforms state-of-the-art static or heuristic-based sparse attention methods, while also being more versatile and flexible to adapt to varying context lengths and sparsity ratios. When applied to long-context fine-tuning with YaRN, SeerAttention can achieve a remarkable 90% sparsity ratio at a 32k context length with minimal perplexity loss, offering a 5.67x speedup over FlashAttention-2.

Large Language Models (LLMs) have exhibited significant potential in performing diverse tasks, including the ability to call functions or use external tools to enhance their performance. While current research on function calling by LLMs primarily focuses on single-turn interactions, this paper addresses the overlooked necessity for LLMs to engage in multi-turn function calling--critical for handling compositional, real-world queries that require planning with functions but not only use functions. To facilitate this, we introduce an approach, BUTTON, which generates synthetic compositional instruction tuning data via bottom-up instruction construction and top-down trajectory generation. In the bottom-up phase, we generate simple atomic tasks based on real-world scenarios and build compositional tasks using heuristic strategies based on atomic tasks. Corresponding functions are then developed for these compositional tasks. The top-down phase features a multi-agent environment where interactions among simulated humans, assistants, and tools are utilized to gather multi-turn function calling trajectories. This approach ensures task compositionality and allows for effective function and trajectory generation by examining atomic tasks within compositional tasks. We produce a dataset BUTTONInstruct comprising 8k data points and demonstrate its effectiveness through extensive experiments across various LLMs.

Vision-language pre-training (VLP) models, trained on large-scale image-text pairs, have become widely used across a variety of downstream vision-and-language (V+L) tasks. This widespread adoption raises concerns about their vulnerability to adversarial attacks. Non-universal adversarial attacks, while effective, are often impractical for real-time online applications due to their high computational demands per data instance. Recently, universal adversarial perturbations (UAPs) have been introduced as a solution, but existing generator-based UAP methods are significantly time-consuming. To overcome the limitation, we propose a direct optimization-based UAP approach, termed DO-UAP, which significantly reduces resource consumption while maintaining high attack performance. Specifically, we explore the necessity of multimodal loss design and introduce a useful data augmentation strategy. Extensive experiments conducted on three benchmark VLP datasets, six popular VLP models, and three classical downstream tasks demonstrate the efficiency and effectiveness of DO-UAP. Specifically, our approach drastically decreases the time consumption by 23-fold while achieving a better attack performance., Comment: 11 pages

Understanding the spatial distribution of palms within tropical forests is essential for effective ecological monitoring, conservation strategies, and the sustainable integration of natural forest products into local and global supply chains. However, the analysis of remotely sensed data in these environments faces significant challenges, such as overlapping palm and tree crowns, uneven shading across the canopy surface, and the heterogeneous nature of the forest landscapes, which often affect the performance of palm detection and segmentation algorithms. To overcome these issues, we introduce PalmDSNet, a deep learning framework for real-time detection, segmentation, and counting of canopy palms. Additionally, we employ a bimodal reproduction algorithm that simulates palm spatial propagation to further enhance the understanding of these point patterns using PalmDSNet's results. We used UAV-captured imagery to create orthomosaics from 21 sites across western Ecuadorian tropical forests, covering a gradient from the everwet Choc\'o forests near Colombia to the drier forests of southwestern Ecuador. Expert annotations were used to create a comprehensive dataset, including 7,356 bounding boxes on image patches and 7,603 palm centers across five orthomosaics, encompassing a total area of 449 hectares. By combining PalmDSNet with the bimodal reproduction algorithm, which optimizes parameters for both local and global spatial variability, we effectively simulate the spatial distribution of palms in diverse and dense tropical environments, validating its utility for advanced applications in tropical forest monitoring and remote sensing analysis., Comment: 25 pages, 8 figures, 5 tables

The recent breakthrough of realizing the Bose-Einstein condensate of polar molecules and degenerate Fermi molecules in three dimensions relies crucially on the microwave shielding technique, which strongly suppresses the collision loss between molecules. In this letter, we show that the cooperation of microwave shielding and dipolar interaction naturally leads to the emergence of a synthetic gauge field. Unlike that studied in cold atoms before, this gauge field couples to the relative motion of every two molecules instead of single-particle motion, therefore being a mutual gauge field. In this case, every molecule carrying a synthetic charge sees the other molecule as carrying the source of the magnetic field, and the spatial distribution of the magnetic field is reminiscent of a solenoid attached to the molecule. In other words, in addition to microwave-shielded interaction, another part of the interaction between two molecules behaves as a charge interacting with a solenoid, which was missed in the previous discussion. We argue that the physical manifestation of this gauge field is breaking time-reversal symmetry in the collective spatial motion of molecules. Finally, we discuss the challenges in quantitatively studying such a quantum many-body system., Comment: 6 pages, 4 figures

Human feedback is crucial in the interactions between humans and Large Language Models (LLMs). However, existing research primarily focuses on benchmarking LLMs in single-turn dialogues. Even in benchmarks designed for multi-turn dialogues, the user inputs are often independent, neglecting the nuanced and complex nature of human feedback within real-world usage scenarios. To fill this research gap, we introduce FB-Bench, a fine-grained, multi-task benchmark designed to evaluate LLMs' responsiveness to human feedback in real-world usage scenarios. Drawing from the two main interaction scenarios, FB-Bench comprises 734 meticulously curated samples, encompassing eight task types, five deficiency types of response, and nine feedback types. We extensively evaluate a broad array of popular LLMs, revealing significant variations in their performance across different interaction scenarios. Further analysis indicates that task, human feedback, and deficiencies of previous responses can also significantly impact LLMs' responsiveness. Our findings underscore both the strengths and limitations of current models, providing valuable insights and directions for future research. Both the toolkits and the dataset of FB-Bench are available at https://github.com/PKU-Baichuan-MLSystemLab/FB-Bench.

The salient multimodal capabilities and interactive experience of GPT-4o highlight its critical role in practical applications, yet it lacks a high-performing open-source counterpart. In this paper, we introduce Baichuan-Omni, the first open-source 7B Multimodal Large Language Model (MLLM) adept at concurrently processing and analyzing modalities of image, video, audio, and text, while delivering an advanced multimodal interactive experience and strong performance. We propose an effective multimodal training schema starting with 7B model and proceeding through two stages of multimodal alignment and multitask fine-tuning across audio, image, video, and text modal. This approach equips the language model with the ability to handle visual and audio data effectively. Demonstrating strong performance across various omni-modal and multimodal benchmarks, we aim for this contribution to serve as a competitive baseline for the open-source community in advancing multimodal understanding and real-time interaction.

The network is becoming Domain-Specific, which requires on-demand design of the network protocols, as well as the microarchitecture of the NIC. However, to develop such a NIC is not that easy. Since the scissor gap between network speed and the growth of CPU frequency is expanding, most of the protocols need to be offloaded to hardware. The process of designing, verifying and optimizing a domain-specific NIC usually takes great effort, which hinders the rapid iteration of new protocols and algorithms. In this paper, we propose JingZhao, an open-sourced framework for NIC prototyping, which could be leveraged to rapidly implement a domain-specific NIC. JingZhao provides several building blocks, as well as a full-fledged RDMA NIC, to help rapidly prototype a high-performance NIC. Our evaluation results show that new network functions can be easily integrated into the framework, and achieve line-rate packet processing., Comment: 12 pages. 14 figures

The languages of logics based on team semantics typically only allow atomic negation or restricted negation. In this paper, we explore propositional team-based logics with full (intuitionistic) negation. We demonstrate that including full intutionistic negation does not complicate the axiomatization of propositional team-based logics with the downward closure property. We also review known expressive completeness results for these logics, highlighting how relevant complemented properties are expressed in propositional dependence logic without directly using negation. Building on these insights, we also prove a new result: propositional logic extended with both dependence and inclusion atoms is expressively complete.

While in-context learning with large language models (LLMs) has shown impressive performance, we have discovered a unique miscalibration behavior where both correct and incorrect predictions are assigned the same level of confidence. We refer to this phenomenon as indiscriminate miscalibration. We found that traditional calibration metrics, such as Expected Calibrated Errors (ECEs), are unable to capture this behavior effectively. To address this issue, we propose new metrics to measure the severity of indiscriminate miscalibration. Additionally, we develop a novel in-context comparative inference method to alleviate miscalibrations and improve classification performance. Through extensive experiments on five datasets, we demonstrate that our proposed method can achieve more accurate and calibrated predictions compared to regular zero-shot and few-shot prompting., Comment: 19 pages

Planning contact-rich interactions for multi-finger manipulation is challenging due to the high-dimensionality and hybrid nature of dynamics. Recent advances in data-driven methods have shown promise, but are sensitive to the quality of training data. Combining learning with classical methods like trajectory optimization and search adds additional structure to the problem and domain knowledge in the form of constraints, which can lead to outperforming the data on which models are trained. We present Diffusion-Informed Probabilistic Contact Search (DIPS), which uses an A* search to plan a sequence of contact modes informed by a diffusion model. We train the diffusion model on a dataset of demonstrations consisting of contact modes and trajectories generated by a trajectory optimizer given those modes. In addition, we use a particle filter-inspired method to reason about variability in diffusion sampling arising from model error, estimating likelihoods of trajectories using a learned discriminator. We show that our method outperforms ablations that do not reason about variability and can plan contact sequences that outperform those found in training data across multiple tasks. We evaluate on simulated tabletop card sliding and screwdriver turning tasks, as well as the screwdriver task in hardware to show that our combined learning and planning approach transfers to the real world.

Probing learned concepts in large language models (LLMs) is crucial for understanding how semantic knowledge is encoded internally. Training linear classifiers on probing tasks is a principle approach to denote the vector of a certain concept in the representation space. However, the single vector identified for a concept varies with both data and training, making it less robust and weakening its effectiveness in real-world applications. To address this challenge, we propose an approach to approximate the subspace representing a specific concept. Built on linear probing classifiers, we extend the concept vectors into Gaussian Concept Subspace (GCS). We demonstrate GCS's effectiveness through measuring its faithfulness and plausibility across multiple LLMs with different sizes and architectures. Additionally, we use representation intervention tasks to showcase its efficacy in real-world applications such as emotion steering. Experimental results indicate that GCS concept vectors have the potential to balance steering performance and maintaining the fluency in natural language generation tasks., Comment: 28 pages, 9 figures

Autonomous racing has rapidly gained research attention. Traditionally, racing cars rely on 2D LiDAR as their primary visual system. In this work, we explore the integration of an event camera with the existing system to provide enhanced temporal information. Our goal is to fuse the 2D LiDAR data with event data in an end-to-end learning framework for steering prediction, which is crucial for autonomous racing. To the best of our knowledge, this is the first study addressing this challenging research topic. We start by creating a multisensor dataset specifically for steering prediction. Using this dataset, we establish a benchmark by evaluating various SOTA fusion methods. Our observations reveal that existing methods often incur substantial computational costs. To address this, we apply low-rank techniques to propose a novel, efficient, and effective fusion design. We introduce a new fusion learning policy to guide the fusion process, enhancing robustness against misalignment. Our fusion architecture provides better steering prediction than LiDAR alone, significantly reducing the RMSE from 7.72 to 1.28. Compared to the second-best fusion method, our work represents only 11% of the learnable parameters while achieving better accuracy. The source code, dataset, and benchmark will be released to promote future research.

The exotic quantum states emerging in the quasicrystal (QC) have attracted extensive interest because of various properties absent in the crystal. In this paper, we systematically study the Holstein model at half filling on a prototypical structure of QC, namely rhombic Penrose lattice, aiming at investigating the superconductivity (SC) and other intertwined ordering arising from the interplay between quasiperiodicity and electron-phonon ({\it e}-ph) interaction. Through unbiased sign-problem-free determinant quantum Monte Carlo simulations, we reveal the salient features of the ground-state phase diagram. Distinct from the results on bipartite periodic lattices at half filling, SC is dominant in a large parameter regime on the Penrose lattice. When {\it e}-ph coupling is sufficiently strong, charge-density-wave order appears and strongly suppresses the SC. The strongest SC emerges at intermediate {\it e}-ph coupling strength and pronounced pairing fluctuation exists above the SC transition temperature. The strong pairing originates from the cooperative effects of unique lattice structure and macroscopically degenerate confined states at Fermi energy which uniquely exist on the Penrose lattice. Moreover, we demonstrate the forbidden ladders substantially suppress the phase coherence of SC. Our unbiased numerical results suggest that Penrose lattice is a potential platform to realize strong SC pairing, providing a promising avenue to searching for relatively high-$T_c$ SC dominantly induced by {\it e}-ph coupling., Comment: 11 pages and 10 figures

We explore self-induced parametric coupling, also called internal resonances (IRs), in a membrane nanoelectromechanical system. Specifically, we focus on the formation of a limit cycle manifesting as a phononic frequency comb. Utilizing a pump-noisy-probe technique and theoretical modeling, we reveal the behavior of mechanical excitations revealing themselves as sidebands of the stationary IR response. We find that when the energy-absorbing excitation of a lower mode is parametrically-upconverted to hybridize with a higher mode, significant squeezing and bimodality in the upper mode occurs. Instead, when the upconverted absorbing excitation hybridizes with an emitting sideband of the higher mode, a Hopf bifurcation occurs and a limit cycle forms, manifesting as a frequency comb. We thus reveal a unique mechanism to obtain frequency combs in parametrically-coupled modes. We furthermore demonstrate a rich variety of IR effects, the origin of which significantly extends beyond standard linear parametric coupling phenomena. Our findings enhance the understanding of energy transfer mechanisms with implications for advanced sensing technologies and novel phononic metamaterials.

The properties displayed by graphene at van Hove singularities (VHS) have caught significant attention in recent years. The emergence of exotic quantum states at these singularities prompts investigations on their evolution within the realm of multilayer stacking structures. In our research, we delve into the study of a repulsive Hubbard model focusing on the AA-stacked bilayer graphene at VHS. Within the system's ground state, each of the top and bottom layers hosts a set of spin-density waves (SDWs). These SDWs each takes on three mutually perpendicular spin polarization directions. Importantly, there is noteworthy feature that their spin polarization directions in the two layers exist as elegant embodiments of antiferromagnetic arrangement, persvading the structure with a striking pattern. Referred to in prior research as the chiral SDWs, this intralayer density wave structure confers the system the characteristics of a Chern topological insulator. However, what is particularly fascinating is the pure divergence of the bilayer structure's topological traits when compared to its monolayer counterpart. The system exhibits a profound symmetry known as $Z_2$, preserving its invariance under the combined operations of time-reversal and interlayer exchange. Consequentely, the system's ground state manifests a seemingly trivial Chern number, yet harbors a profound and intricate nontrivial $Z_2$ topological invariant. These remarkable observations align our findings with the conceptual framework of the quantum spin Hall effect.

High-fidelity direct numerical simulation of turbulent flows for most real-world applications remains an outstanding computational challenge. Several machine learning approaches have recently been proposed to alleviate the computational cost even though they become unstable or unphysical for long time predictions. We identify that the Fourier neural operator (FNO) based models combined with a partial differential equation (PDE) solver can accelerate fluid dynamic simulations and thus address computational expense of large-scale turbulence simulations. We treat the FNO model on the same footing as a PDE solver and answer important questions about the volume and temporal resolution of data required to build pre-trained models for turbulence. We also discuss the pitfalls of purely data-driven approaches that need to be avoided by the machine learning models to become viable and competitive tools for long time simulations of turbulence.

Generative AI has shown its values for many software engineering tasks. Still in its infancy, large language model (LLM)-based proof generation lags behind LLM-based code generation. In this paper, we present AutoVerus. AutoVerus uses LLM to automatically generate correctness proof for Rust code. AutoVerus is designed to match the unique features of Verus, a verification tool that can prove the correctness of Rust code using proofs and specifications also written in Rust. AutoVerus consists of a network of LLM agents that are crafted and orchestrated to mimic human experts' three phases of proof construction: preliminary proof generation, proof refinement guided by generic tips, and proof debugging guided by verification errors. To thoroughly evaluate AutoVerus and help foster future research in this direction, we have built a benchmark suite of 150 non-trivial proof tasks, based on existing code-generation benchmarks and verification benchmarks. Our evaluation shows that AutoVerus can automatically generate correct proof for more than 90% of them, with more than half of them tackled in less than 30 seconds or 3 LLM calls.

Video generation using diffusion-based models is constrained by high computational costs due to the frame-wise iterative diffusion process. This work presents a Diffusion Reuse MOtion (Dr. Mo) network to accelerate latent video generation. Our key discovery is that coarse-grained noises in earlier denoising steps have demonstrated high motion consistency across consecutive video frames. Following this observation, Dr. Mo propagates those coarse-grained noises onto the next frame by incorporating carefully designed, lightweight inter-frame motions, eliminating massive computational redundancy in frame-wise diffusion models. The more sensitive and fine-grained noises are still acquired via later denoising steps, which can be essential to retain visual qualities. As such, deciding which intermediate steps should switch from motion-based propagations to denoising can be a crucial problem and a key tradeoff between efficiency and quality. Dr. Mo employs a meta-network named Denoising Step Selector (DSS) to dynamically determine desirable intermediate steps across video frames. Extensive evaluations on video generation and editing tasks have shown that Dr. Mo can substantially accelerate diffusion models in video tasks with improved visual qualities.

Transformer-based Large Language Models (LLMs) have become increasingly important. However, due to the quadratic time complexity of attention computation, scaling LLMs to longer contexts incurs extremely slow inference latency and high GPU memory consumption for caching key-value (KV) vectors. This paper proposes RetrievalAttention, a training-free approach to both accelerate attention computation and reduce GPU memory consumption. By leveraging the dynamic sparsity of attention mechanism, RetrievalAttention proposes to use approximate nearest neighbor search (ANNS) indexes for KV vectors in CPU memory and retrieves the most relevant ones with vector search during generation. Unfortunately, we observe that the off-the-shelf ANNS indexes are often ineffective for such retrieval tasks due to the out-of-distribution (OOD) between query vectors and key vectors in attention mechanism. RetrievalAttention addresses the OOD challenge by designing an attention-aware vector search algorithm that can adapt to the distribution of query vectors. Our evaluation shows that RetrievalAttention only needs to access 1--3% of data while maintaining high model accuracy. This leads to significant reduction in the inference cost of long-context LLMs with much lower GPU memory footprint. In particular, RetrievalAttention only needs a single NVIDIA RTX4090 (24GB) for serving 128K tokens in LLMs with 8B parameters, which is capable of generating one token in 0.188 seconds., Comment: 16 pages

Recent advances in machine learning (ML) for automating analog circuit synthesis have been significant, yet challenges remain. A critical gap is the lack of a standardized evaluation framework, compounded by various process design kits (PDKs), simulation tools, and a limited variety of circuit topologies. These factors hinder direct comparisons and the validation of algorithms. To address these shortcomings, we introduced AnalogGym, an open-source testing suite designed to provide fair and comprehensive evaluations. AnalogGym includes 30 circuit topologies in five categories: sensing front ends, voltage references, low dropout regulators, amplifiers, and phase-locked loops. It supports several technology nodes for academic and commercial applications and is compatible with commercial simulators such as Cadence Spectre, Synopsys HSPICE, and the open-source simulator Ngspice. AnalogGym standardizes the assessment of ML algorithms in analog circuit synthesis and promotes reproducibility with its open datasets and detailed benchmark specifications. AnalogGym's user-friendly design allows researchers to easily adapt it for robust, transparent comparisons of state-of-the-art methods, while also exposing them to real-world industrial design challenges, enhancing the practical relevance of their work. Additionally, we have conducted a comprehensive comparison study of various analog sizing methods on AnalogGym, highlighting the capabilities and advantages of different approaches. AnalogGym is available in the GitHub repository https://github.com/CODA-Team/AnalogGym. The documentation is also available at http://coda-team.github.io/AnalogGym/.

Hot Jupiters should initially form at considerable distances from host stars and subsequently migrate towards inner regions, supported directly by transit timing variation (TTV). We report the TTV of K2-237b, using reproduced timings fitted from \textit{Kepler} K2 and \textit{TESS} data. The timings span from 2016 to 2021, leading to an observational baseline of 5 years. The timing evolution presents a significant bias to a constant period scenario. The model evidence is evaluated utilizing the Bayesian Information Criterion (BIC), which favours the scenario of period decay with a $\Delta$BIC of 14.1. The detected TTV induces a period decay rate ($\dot{P}$) of -1.14$\pm$0.28$\times$10$^{-8}$ days per day ($-$0.36 s/year). Fitting the spectral energy distribution, we find infrared excess at the significance level of 1.5 $\sigma$ for WISE W1 and W2 bands, and 2 $\sigma$ level for W3 and W4 bands. This potentially reveals the existence of a stellar disk, consisting of hot dust at 800$\pm$300 K, showing a $L_{dust}/L_{\ast}$ of 5$\pm$3$\times$10$^{-3}$. We obtain a stellar age of 1.0$^{+1.4}_{-0.7}$$\times$10$^{9}$ yr from isochrone fitting. The properties of K2-237b potentially serve as a direct observational support to the planet disk migration though more observation are needed., Comment: 5 pages, Accepted for publication in MNRAS Letters

In this paper, we study branching Brownian motion with absorption, in which particles undergo Brownian motions and are killed upon hitting the absorption barrier. We prove that the empirical distribution function of the maximum of this process converges almost surely to a randomly shifted Gumbel distribution., Comment: 14 pages, 3 figures

The high performance of tree ensemble classifiers benefits from a large set of rules, which, in turn, makes the models hard to understand. To improve interpretability, existing methods extract a subset of rules for approximation using model reduction techniques. However, by focusing on the reduced rule set, these methods often lose fidelity and ignore anomalous rules that, despite their infrequency, play crucial roles in real-world applications. This paper introduces a scalable visual analysis method to explain tree ensemble classifiers that contain tens of thousands of rules. The key idea is to address the issue of losing fidelity by adaptively organizing the rules as a hierarchy rather than reducing them. To ensure the inclusion of anomalous rules, we develop an anomaly-biased model reduction method to prioritize these rules at each hierarchical level. Synergized with this hierarchical organization of rules, we develop a matrix-based hierarchical visualization to support exploration at different levels of detail. Our quantitative experiments and case studies demonstrate how our method fosters a deeper understanding of both common and anomalous rules, thereby enhancing interpretability without sacrificing comprehensiveness., Comment: 15 pages, 10 figures

In this work, we propose a deep learning-based imaging method for addressing the multi-frequency electromagnetic (EM) inverse scattering problem (ISP). By combining deep learning technology with EM physical laws, we have successfully developed a multi-frequency neural Born iterative method (NeuralBIM), guided by the principles of the single-frequency NeuralBIM. This method integrates multitask learning techniques with NeuralBIM's efficient iterative inversion process to construct a robust multi-frequency Born iterative inversion model. During training, the model employs a multitask learning approach guided by homoscedastic uncertainty to adaptively allocate the weights of each frequency's data. Additionally, an unsupervised learning method, constrained by the physical laws of ISP, is used to train the multi-frequency NeuralBIM model, eliminating the need for contrast and total field data. The effectiveness of the multi-frequency NeuralBIM is validated through synthetic and experimental data, demonstrating improvements in accuracy and computational efficiency for solving ISP. Moreover, this method exhibits strong generalization capabilities and noise resistance. The multi-frequency NeuralBIM method explores a novel inversion method for multi-frequency EM data and provides an effective solution for the electromagnetic ISP of multi-frequency data.

Variational continual learning (VCL) is a turn-key learning algorithm that has state-of-the-art performance among the best continual learning models. In our work, we explore an extension of the generalized variational continual learning (GVCL) model, named AutoVCL, which combines task heuristics for informed learning and model optimization. We demonstrate that our model outperforms the standard GVCL with fixed hyperparameters, benefiting from the automatic adjustment of the hyperparameter based on the difficulty and similarity of the incoming task compared to the previous tasks., Comment: 4 pages, 2 figures, 3 tables

The general capabilities of Large Language Models (LLM) highly rely on the composition and selection on extensive pretraining datasets, treated as commercial secrets by several institutions. To mitigate this issue, we open-source the details of a universally applicable data processing pipeline and validate its effectiveness and potential by introducing a competitive LLM baseline. Specifically, the data processing pipeline consists of broad collection to scale up and reweighting to improve quality. We then pretrain a 7B model BaichuanSEED with 3T tokens processed by our pipeline without any deliberate downstream task-related optimization, followed by an easy but effective supervised fine-tuning stage. BaichuanSEED demonstrates consistency and predictability throughout training and achieves comparable performance on comprehensive benchmarks with several commercial advanced large language models, such as Qwen1.5 and Llama3. We also conduct several heuristic experiments to discuss the potential for further optimization of downstream tasks, such as mathematics and coding., Comment: 19 pages, 6 figures

Parameterizing finger rolling and finger-object contacts in a differentiable manner is important for formulating dexterous manipulation as a trajectory optimization problem. In contrast to previous methods which often assume simplified geometries of the robot and object or do not explicitly model finger rolling, we propose a method to further extend the capabilities of dexterous manipulation by accounting for non-trivial geometries of both the robot and the object. By integrating the object's Signed Distance Field (SDF) with a sampling method, our method estimates contact and rolling-related variables and includes those in a trajectory optimization framework. This formulation naturally allows for the emergence of finger-rolling behaviors, enabling the robot to locally adjust the contact points. Our method is tested in a peg alignment task and a screwdriver turning task, where it outperforms the baselines in terms of achieving desired object configurations and avoiding dropping the object. We also successfully apply our method to a real-world screwdriver turning task, demonstrating its robustness to the sim2real gap.

Quantum money is the first invention in quantum information science, promising advantages over classical money by simultaneously achieving unforgeability, user privacy, and instant validation. However, standard quantum money relies on quantum memories and long-distance quantum communication, which are technologically extremely challenging. Quantum "S-money" tokens eliminate these technological requirements while preserving unforgeability, user privacy, and instant validation. Here, we report the first full experimental demonstration of quantum S-tokens, proven secure despite errors, losses and experimental imperfections. The heralded single-photon source with a high system efficiency of 88.24% protects against arbitrary multi-photon attacks arising from losses in the quantum token generation. Following short-range quantum communication, the token is stored, transacted, and verified using classical bits. We demonstrate a transaction time advantage over intra-city 2.77 km and inter-city 60.54 km optical fibre networks, compared with optimal classical cross-checking schemes. Our implementation demonstrates the practicality of quantum S-tokens for applications requiring high security, privacy and minimal transaction times, like financial trading and network control. It is also the first demonstration of a quantitative quantum time advantage in relativistic cryptography, showing the enhanced cryptographic power of simultaneously considering quantum and relativistic physics., Comment: 74 pages, 6 figures

We show that for a family of quantum walk models with electric fields, the spectrum is the unit circle for any irrational field. The result also holds for the associated CMV matrices defined by skew-shifts. Generalizations to CMV matrices with skew-shifts on higher dimensional torus are also obtained.

Artisanal and Small-Scale Gold Mining (ASGM) is a low-cost yet highly destructive mining practice, leading to environmental disasters across the world's tropical watersheds. The topic of ASGM spans multiple domains of research and information, including natural and social systems, and knowledge is often atomized across a diversity of media and documents. We therefore introduce a knowledge graph (ASGM-KG) that consolidates and provides crucial information about ASGM practices and their environmental effects. The current version of ASGM-KG consists of 1,899 triples extracted using a large language model (LLM) from documents and reports published by both non-governmental and governmental organizations. These documents were carefully selected by a group of tropical ecologists with expertise in ASGM. This knowledge graph was validated using two methods. First, a small team of ASGM experts reviewed and labeled triples as factual or non-factual. Second, we devised and applied an automated factual reduction framework that relies on a search engine and an LLM for labeling triples. Our framework performs as well as five baselines on a publicly available knowledge graph and achieves over 90 accuracy on our ASGM-KG validated by domain experts. ASGM-KG demonstrates an advancement in knowledge aggregation and representation for complex, interdisciplinary environmental crises such as ASGM.

A typical VLSI design flow is divided into separated front-end logic synthesis and back-end physical design (PD) stages, which often require costly iterations between these stages to achieve design closure. Existing approaches face significant challenges, notably in utilizing feedback from physical metrics to better adapt and refine synthesis operations, and in establishing a unified and comprehensive metric. This paper introduces a new Primitive logic gate placement guided technology MAPping (PigMAP) framework to address these challenges. With approximating technology-independent spatial information, we develop a novel wirelength (WL) driven mapping algorithm to produce PD-friendly netlists. PigMAP is equipped with two schemes: a performance mode that focuses on optimizing the critical path WL to achieve high performance, and a power mode that aims to minimize the total WL, resulting in balanced power and performance outcomes. We evaluate our framework using the EPFL benchmark suites with ASAP7 technology, using the OpenROAD tool for place-and-route. Compared with OpenROAD flow scripts, performance mode reduces delay by 14% while increasing power consumption by only 6%. Meanwhile, power mode achieves a 3% improvement in delay and a 9% reduction in power consumption., Comment: 9 pages, 8 figures, 2 tables

Recent advancements in autonomous driving, augmented reality, robotics, and embodied intelligence have necessitated 3D perception algorithms. However, current 3D perception methods, particularly small models, struggle with processing logical reasoning, question-answering, and handling open scenario categories. On the other hand, generative multimodal large language models (MLLMs) excel in general capacity but underperform in 3D tasks, due to weak spatial and local object perception, poor text-based geometric numerical output, and inability to handle camera focal variations. To address these challenges, we propose the following solutions: Spatial-Enhanced Local Feature Mining for better spatial feature extraction, 3D Query Token-Derived Info Decoding for precise geometric regression, and Geometry Projection-Based 3D Reasoning for handling camera focal length variations. We employ parameter-efficient fine-tuning for a pre-trained MLLM and develop LLMI3D, a powerful 3D perception MLLM. Additionally, we have constructed the IG3D dataset, which provides fine-grained descriptions and question-answer annotations. Extensive experiments demonstrate that our LLMI3D achieves state-of-the-art performance, significantly outperforming existing methods.

This paper introduces rStar, a self-play mutual reasoning approach that significantly improves reasoning capabilities of small language models (SLMs) without fine-tuning or superior models. rStar decouples reasoning into a self-play mutual generation-discrimination process. First, a target SLM augments the Monte Carlo Tree Search (MCTS) with a rich set of human-like reasoning actions to construct higher quality reasoning trajectories. Next, another SLM, with capabilities similar to the target SLM, acts as a discriminator to verify each trajectory generated by the target SLM. The mutually agreed reasoning trajectories are considered mutual consistent, thus are more likely to be correct. Extensive experiments across five SLMs demonstrate rStar can effectively solve diverse reasoning problems, including GSM8K, GSM-Hard, MATH, SVAMP, and StrategyQA. Remarkably, rStar boosts GSM8K accuracy from 12.51% to 63.91% for LLaMA2-7B, from 36.46% to 81.88% for Mistral-7B, from 74.53% to 91.13% for LLaMA3-8B-Instruct. Code will be available at https://github.com/zhentingqi/rStar.

As large language model (LLM) inference demands ever-greater resources, there is a rapid growing trend of using low-bit weights to shrink memory usage and boost inference efficiency. However, these low-bit LLMs introduce the need for mixed-precision matrix multiplication (mpGEMM), which is a crucial yet under-explored operation that involves multiplying lower-precision weights with higher-precision activations. Unfortunately, current hardware does not natively support mpGEMM, resulting in indirect and inefficient dequantization-based implementations. To address the mpGEMM requirements in low-bit LLMs, we explored the lookup table (LUT)-based approach for mpGEMM. However, a conventional LUT implementation falls short of its potential. To fully harness the power of LUT-based mpGEMM, we introduce LUT Tensor Core, a software-hardware co-design optimized for low-bit LLM inference. Specifically, we introduce software-based operator fusion and table symmetrization techniques to optimize table precompute and table storage, respectively. Then, LUT Tensor Core proposes the hardware design featuring an elongated tiling shape design to enhance table reuse and a bit-serial design to support various precision combinations in mpGEMM. Moreover, we design an end-to-end compilation stack with new instructions for LUT-based mpGEMM, enabling efficient LLM compilation and optimizations. The evaluation on low-bit LLMs (e.g., BitNet, LLAMA) shows that LUT Tensor Core achieves more than a magnitude of improvements on both compute density and energy efficiency.