Search

We introduce Baichuan-Omni-1.5, an omni-modal model that not only has omni-modal understanding capabilities but also provides end-to-end audio generation capabilities. To achieve fluent and high-quality interaction across modalities without compromising the capabilities of any modality, we prioritized optimizing three key aspects. First, we establish a comprehensive data cleaning and synthesis pipeline for multimodal data, obtaining about 500B high-quality data (text, audio, and vision). Second, an audio-tokenizer (Baichuan-Audio-Tokenizer) has been designed to capture both semantic and acoustic information from audio, enabling seamless integration and enhanced compatibility with MLLM. Lastly, we designed a multi-stage training strategy that progressively integrates multimodal alignment and multitask fine-tuning, ensuring effective synergy across all modalities. Baichuan-Omni-1.5 leads contemporary models (including GPT4o-mini and MiniCPM-o 2.6) in terms of comprehensive omni-modal capabilities. Notably, it achieves results comparable to leading models such as Qwen2-VL-72B across various multimodal medical benchmarks.

Recent advances in Large Language Models (LLMs) have upgraded them from sophisticated text generators to autonomous agents capable of corporation and tool use in multi-agent systems (MASs). However, the robustness of these LLM-based MASs, especially under knowledge conflicts, remains unclear. In this paper, we design four comprehensive metrics to investigate the robustness of MASs when facing mild or task-critical knowledge conflicts. We first analyze mild knowledge conflicts introduced by heterogeneous agents and find that they do not harm system robustness but instead improve collaborative decision-making. Next, we investigate task-critical knowledge conflicts by synthesizing knowledge conflicts and embedding them into one of the agents. Our results show that these conflicts have surprisingly little to no impact on MAS robustness. Furthermore, we observe that MASs demonstrate certain self-repairing capabilities by reducing their reliance on knowledge conflicts and adopting alternative solution paths to maintain stability. Finally, we conduct ablation studies on the knowledge conflict number, agent number, and interaction rounds, finding that the self-repairing capability of MASs has intrinsic limits, and all findings hold consistently across various factors. Our code is publicly available at https://github.com/wbw625/MultiAgentRobustness., Comment: Working in progress

Non-Abelian topological order (TO) enables topologically protected quantum computation with its anyonic quasiparticles. Recently, TO with $S_3$ gauge symmetry was identified as a sweet spot -- simple enough to emerge from finite-depth adaptive circuits yet powerful enough to support a universal topological gate-set. In these notes, we review how anyon braiding and measurement in $S_3$ TO are primitives for topological quantum computation and we explicitly demonstrate universality. These topological operations are made concrete in the $S_3$ quantum double lattice model, aided by the introduction of a generalized ribbon operator. This provides a roadmap for near-term quantum platforms., Comment: 70 pages, 24 figures

Moir\'e heterobilayers exhibiting spatially varying band alignment and electron and hole localization that can be precisely controlled through the twist angle have emerged as exciting platforms for studying complex quantum phenomena. While most heterobilayers of transition metal dichalcogenides (TMDs) have a type II band alignment, the introduction of type I band alignment could enable stronger light-matter coupling and enhanced radiative emission. Here, we show through a combination of first-principles GW plus Bethe Salpeter equation (GW-BSE) calculations and time- and angle-resolved photoemission spectroscopy (tr-ARPES) measurements that contrary to previous understanding, the MoSe2/WS2 heterobilayer has a type I band alignment at large twist angles and simultaneous regions of type I and type II band alignment due to the structural reconstruction in different high symmetry regions at small twist angles. In tr-ARPES, consistent with our calculations, a long-lived electron population is only observed in MoSe2 for samples with large twist angles, while in samples with small twist angles, signals from two distinct long-lived excitons are observed. Moreover, despite the near degeneracy of the conduction bands of the two layers, no excitonic hybridization occurs, suggesting that previously observed absorption peaks in this material arise from lattice reconstruction. Our findings clarify the complex energy landscape in MoSe2/WS2 heterostructures, where the coexistence of type I and type II band alignment opens the door to moir\'e-tunable optoelectronic devices with intrinsic lateral heterojunctions.

Myopia, projected to affect 50% population globally by 2050, is a leading cause of vision loss. Eyes with pathological myopia exhibit distinctive shape distributions, which are closely linked to the progression of vision-threatening complications. Recent understanding of eye-shape-based biomarkers requires magnetic resonance imaging (MRI), however, it is costly and unrealistic in routine ophthalmology clinics. We present Fundus2Globe, the first AI framework that synthesizes patient-specific 3D eye globes from ubiquitous 2D color fundus photographs (CFPs) and routine metadata (axial length, spherical equivalent), bypassing MRI dependency. By integrating a 3D morphable eye model (encoding biomechanical shape priors) with a latent diffusion model, our approach achieves submillimeter accuracy in reconstructing posterior ocular anatomy efficiently. Fundus2Globe uniquely quantifies how vision-threatening lesions (e.g., staphylomas) in CFPs correlate with MRI-validated 3D shape abnormalities, enabling clinicians to simulate posterior segment changes in response to refractive shifts. External validation demonstrates its robust generation performance, ensuring fairness across underrepresented groups. By transforming 2D fundus imaging into 3D digital replicas of ocular structures, Fundus2Globe is a gateway for precision ophthalmology, laying the foundation for AI-driven, personalized myopia management., Comment: 24 pages, 6 figures

The spatio-temporal relationship between the pixels of a video carries critical information for low-level 4D perception. A single model that reasons about it should be able to solve several such tasks well. Yet, most state-of-the-art methods rely on architectures specialized for the task at hand. We present L4P (pronounced "LAP"), a feedforward, general-purpose architecture that solves low-level 4D perception tasks in a unified framework. L4P combines a ViT-based backbone with per-task heads that are lightweight and therefore do not require extensive training. Despite its general and feedforward formulation, our method matches or surpasses the performance of existing specialized methods on both dense tasks, such as depth or optical flow estimation, and sparse tasks, such as 2D/3D tracking. Moreover, it solves all those tasks at once in a time comparable to that of individual single-task methods.

The vibrational behavior of molecules serves as a crucial fingerprint of their structure, chemical state, and surrounding environment. Neutron vibrational spectroscopy provides comprehensive measurements of vibrational modes without selection rule restrictions. However, analyzing and interpreting the resulting spectra remains a computationally formidable task. Here, we introduce a symmetry-aware neural network that directly predicts Hessian matrices from molecular structures, thereby enabling rapid vibrational spectral reconstruction. Unlike traditional approaches that focus on eigenvalue prediction, the Hessian matrix provides richer, more fundamental information with broader applications and superior extrapolation. This approach also paves the way for predicting other properties, such as reaction pathways. Trained on small molecules, our model achieves spectroscopic-level accuracy, allowing real-time, unambiguous peak assignment. Moreover, it maintains high accuracy for larger molecules, demonstrating strong transferability. This adaptability unlocks new capabilities, including on-the-fly spectral interpretation for future autonomous laboratories, and offers insights into molecular design for targeted chemical pathways.

Unsafe prompts pose significant safety risks to large language models (LLMs). Existing methods for detecting unsafe prompts rely on data-driven fine-tuning to train guardrail models, necessitating significant data and computational resources. In contrast, recent few-shot gradient-based methods emerge, requiring only few safe and unsafe reference prompts. A gradient-based approach identifies unsafe prompts by analyzing consistent patterns of the gradients of safety-critical parameters in LLMs. Although effective, its restriction to directional similarity (cosine similarity) introduces ``directional bias'', limiting its capability to identify unsafe prompts. To overcome this limitation, we introduce GradCoo, a novel gradient co-occurrence analysis method that expands the scope of safety-critical parameter identification to include unsigned gradient similarity, thereby reducing the impact of ``directional bias'' and enhancing the accuracy of unsafe prompt detection. Comprehensive experiments on the widely-used benchmark datasets ToxicChat and XStest demonstrate that our proposed method can achieve state-of-the-art (SOTA) performance compared to existing methods. Moreover, we confirm the generalizability of GradCoo in detecting unsafe prompts across a range of LLM base models with various sizes and origins.

The rapid development of Multimodal Large Language Models (MLLMs) has enabled the integration of multiple modalities, including texts and images, within the large language model (LLM) framework. However, texts and images are usually interconnected, forming a multimodal attributed graph (MMAG). It is underexplored how MLLMs can incorporate the relational information (\textit{i.e.}, graph structure) and semantic information (\textit{i.e.,} texts and images) on such graphs for multimodal comprehension and generation. In this paper, we propose GraphGPT-o, which supports omni-multimodal understanding and creation on MMAGs. We first comprehensively study linearization variants to transform semantic and structural information as input for MLLMs. Then, we propose a hierarchical aligner that enables deep graph encoding, bridging the gap between MMAGs and MLLMs. Finally, we explore the inference choices, adapting MLLM to interleaved text and image generation in graph scenarios. Extensive experiments on three datasets from different domains demonstrate the effectiveness of our proposed method. Datasets and codes will be open-sourced upon acceptance.

Although Large Language Models (LLMs) excel in reasoning and generation for language tasks, they are not specifically designed for multimodal challenges. Training Multimodal Large Language Models (MLLMs), however, is resource-intensive and constrained by various training limitations. In this paper, we propose the Modular-based Visual Contrastive Decoding (MVCD) framework to move this obstacle. Our framework leverages LLMs' In-Context Learning (ICL) capability and the proposed visual contrastive-example decoding (CED), specifically tailored for this framework, without requiring any additional training. By converting visual signals into text and focusing on contrastive output distributions during decoding, we can highlight the new information introduced by contextual examples, explore their connections, and avoid over-reliance on prior encoded knowledge. MVCD enhances LLMs' visual perception to make it see and reason over the input visuals. To demonstrate MVCD's effectiveness, we conduct experiments with four LLMs across five question answering datasets. Our results not only show consistent improvement in model accuracy but well explain the effective components inside our decoding strategy. Our code will be available at https://github.com/Pbhgit/MVCD., Comment: Accepted to ICASSP 2025

This work focuses on the synergy of physical layer covert transmission and millimeter wave (mmWave) integrated sensing and communication (ISAC) to improve the performance, and enable secure internet of things (IoT) systems. Specifically, we employ a physical layer covert transmission as a prism, which can achieve simultaneously transmitting confidential signals to a covert communication user equipment (UE) in the presence of a warden and regular communication UEs. We design the transmit beamforming to guarantee information transmission security, communication quality-of-service (QoS) and sensing accuracy. By considering two different beamforming architectures, i.e., fully digital beamforming (FDBF) and hybrid beamforming (HBF), an optimal design method and a low-cost beamforming scheme are proposed to address the corresponding problems, respectively. Numerical simulations validate the effectiveness and superiority of the proposed FDBF/HBF algorithms compared with traditional algorithms in terms of information transmission security, communication QoS and target detection performance.

Kitaev quantum spin liquids have attracted significant attention in condensed matter physics over the past decade. To understand their emergent quantum phenomena, high-quality single crystals of substantial size are essential. Here, we report the synthesis of single crystals of the Kitaev quantum spin liquid candidate RuBr3, achieving millimeter-sized crystals through a self-flux method under high pressure and high temperature conditions. The crystals exhibit well-defined cleavage planes with a lustrous appearance. Transport characterizations exhibit a narrow band-gap semiconducting behavior with 0.13 eV and 0.11 eV band-gap in ab plane and along c axis, respectively. Magnetic measurement shows a transition to antiferromagnetic (AFM) state at approximately 29 K both in ab plane and along c axis. Notably, the N\'eel temperature increases to 34 K with an applied magnetic field of up to 7 T in the ab plane, but without any change along c axis. The large size and high quality of RuBr3 single crystals provide a valuable platform for investigating various interactions, particularly the Kitaev interaction, and for elucidating the intrinsic physical properties of Kitaev quantum spin liquids., Comment: Chinese Physics Letters accepted

Solving complex reasoning tasks may involve visual understanding, domain knowledge retrieval, numerical calculation, and multi-step reasoning. Existing methods augment large language models (LLMs) with external tools but are restricted to specialized domains, limited tool types, or require additional training data. In this paper, we introduce OctoTools, a training-free, user-friendly, and easily extensible open-source agentic framework designed to tackle complex reasoning across diverse domains. OctoTools introduces standardized tool cards to encapsulate tool functionality, a planner for both high-level and low-level planning, and an executor to carry out tool usage. We validate OctoTools' generality across 16 diverse tasks (including MathVista, MMLU-Pro, MedQA, and GAIA-Text), achieving substantial average accuracy gains of 9.3% over GPT-4o. Furthermore, OctoTools outperforms AutoGen, GPT-Functions and LangChain by up to 10.6% when given the same set of tools. Through comprehensive analysis and ablations, OctoTools demonstrates advantages in task planning, effective tool usage, and multi-step problem solving., Comment: 89 pages, 18 figures. Project website: https://octotools.github.io/

In specialized fields like the scientific domain, constructing large-scale human-annotated datasets poses a significant challenge due to the need for domain expertise. Recent methods have employed large language models to generate synthetic queries, which serve as proxies for actual user queries. However, they lack control over the content generated, often resulting in incomplete coverage of academic concepts in documents. We introduce Concept Coverage-based Query set Generation (CCQGen) framework, designed to generate a set of queries with comprehensive coverage of the document's concepts. A key distinction of CCQGen is that it adaptively adjusts the generation process based on the previously generated queries. We identify concepts not sufficiently covered by previous queries, and leverage them as conditions for subsequent query generation. This approach guides each new query to complement the previous ones, aiding in a thorough understanding of the document. Extensive experiments demonstrate that CCQGen significantly enhances query quality and retrieval performance., Comment: WSDM 2025

This paper proposes a scalable binary CUR low-rank approximation algorithm that leverages parallel selection of representative rows and columns within a deterministic framework. By employing a blockwise adaptive cross approximation strategy, the algorithm efficiently identifies dominant components in large-scale matrices, thereby reducing computational costs. Numerical experiments on $16,384 \times 16,384$ matrices demonstrate remarkable scalability, with execution time decreasing from $12.37$ seconds using $2$ processes to $1.02$ seconds using $64$ processes. The tests on Hilbert matrices and synthetic low-rank matrices across various sizes demonstrate a near-optimal reconstruction accuracy. These results suggest a potential for practical application in large-scale matrix low-rank approximation.

This work demonstrates that diffusion models can achieve font-controllable multilingual text rendering using just raw images without font label annotations. Visual text rendering remains a significant challenge. While recent methods condition diffusion on glyphs, it is impossible to retrieve exact font annotations from large-scale, real-world datasets, which prevents user-specified font control. To address this, we propose a data-driven solution that integrates the conditional diffusion model with a text segmentation model, utilizing segmentation masks to capture and represent fonts in pixel space in a self-supervised manner, thereby eliminating the need for any ground-truth labels and enabling users to customize text rendering with any multilingual font of their choice. The experiment provides a proof of concept of our algorithm in zero-shot text and font editing across diverse fonts and languages, providing valuable insights for the community and industry toward achieving generalized visual text rendering., Comment: This is preliminary work and code will be released at github.com/bowen-upenn/ControlText

English as a Second Language (ESL) learners often encounter unknown words that hinder their text comprehension. Automatically detecting these words as users read can enable computing systems to provide just-in-time definitions, synonyms, or contextual explanations, thereby helping users learn vocabulary in a natural and seamless manner. This paper presents EyeLingo, a transformer-based machine learning method that predicts the probability of unknown words based on text content and eye gaze trajectory in real time with high accuracy. A 20-participant user study revealed that our method can achieve an accuracy of 97.6%, and an F1-score of 71.1%. We implemented a real-time reading assistance prototype to show the effectiveness of EyeLingo. The user study shows improvement in willingness to use and usefulness compared to baseline methods.

Anomaly detection plays a critical role in modern data-driven applications, from identifying fraudulent transactions and safeguarding network infrastructure to monitoring sensor systems for irregular patterns. Traditional approaches, such as distance, density, or cluster-based methods, face significant challenges when applied to high dimensional tensor data, where complex interdependencies across dimensions amplify noise and computational complexity. To address these limitations, this paper leverages Tensor Chidori pseudoskeleton decomposition within a tensor-robust principal component analysis framework to extract low Tucker rank structure while isolating sparse anomalies, ensuring robustness to anomaly detection. We establish theoretical results regarding convergence, and estimation error, demonstrating the stability and accuracy of the proposed approach. Numerical experiments on real-world spatiotemporal data from New York City taxi trip records validate the superiority of the proposed method in detecting anomalous urban events compared to existing benchmark methods. The results underscore the potential of Tensor Chidori pseudoskeleton decomposition to enhance anomaly detection for large-scale, high-dimensional data.

A robot that learns from demonstrations should not just imitate what it sees -- it should understand the high-level concepts that are being demonstrated and generalize them to new tasks. Bilevel planning is a hierarchical model-based approach where predicates (relational state abstractions) can be leveraged to achieve compositional generalization. However, previous bilevel planning approaches depend on predicates that are either hand-engineered or restricted to very simple forms, limiting their scalability to sophisticated, high-dimensional state spaces. To address this limitation, we present IVNTR, the first bilevel planning approach capable of learning neural predicates directly from demonstrations. Our key innovation is a neuro-symbolic bilevel learning framework that mirrors the structure of bilevel planning. In IVNTR, symbolic learning of the predicate "effects" and neural learning of the predicate "functions" alternate, with each providing guidance for the other. We evaluate IVNTR in six diverse robot planning domains, demonstrating its effectiveness in abstracting various continuous and high-dimensional states. While most existing approaches struggle to generalize (with <35% success rate), our IVNTR achieves an average of 77% success rate on unseen tasks. Additionally, we showcase IVNTR on a mobile manipulator, where it learns to perform real-world mobile manipulation tasks and generalizes to unseen test scenarios that feature new objects, new states, and longer task horizons. Our findings underscore the promise of learning and planning with abstractions as a path towards high-level generalization., Comment: 16 pages, 9 figures

Cultural and language factors significantly influence counseling, but Natural Language Processing research has not yet examined whether the findings of conversational analysis for counseling conducted in English apply to other languages. This paper presents a first step towards this direction. We introduce MIDAS (Motivational Interviewing Dataset in Spanish), a counseling dataset created from public video sources that contains expert annotations for counseling reflections and questions. Using this dataset, we explore language-based differences in counselor behavior in English and Spanish and develop classifiers in monolingual and multilingual settings, demonstrating its applications in counselor behavioral coding tasks., Comment: To appear in NAACL 2025 Main Conference

Chip-based soliton frequency microcombs combine compact size, broad bandwidth, and high coherence, presenting a promising solution for integrated optical telecommunications, precision sensing, and spectroscopy. Recent progress in ferroelectric thin films, particularly thin-film Lithium niobate (LN) and thin-film Lithium tantalate (LT), has significantly advanced electro-optic (EO) modulation and soliton microcombs generation, leveraging their strong third-order nonlinearity and high Pockels coefficients. However, achieving soliton frequency combs in X-cut ferroelectric materials remains challenging due to the competing effects of thermo-optic and photorefractive phenomena. These issues hinder the simultaneous realization of soliton generation and high-speed EO modulation. Here, following the thermal-regulated carrier behaviour and auxiliary-laser-assisted approach, we propose a convenient mechanism to suppress both photorefractive and thermal dragging effect at once, and implement a facile method for soliton formation and its long-term stabilization in integrated X-cut LT microresonators for the first time. The resulting mode-locked states exhibit robust stability against perturbations, enabling new pathways for fully integrated photonic circuits that combine Kerr nonlinearity with high-speed EO functionality., Comment: 8 pages, 5 figures, article

Remote sensing image super-resolution (SR) aims to reconstruct high-resolution remote sensing images from low-resolution inputs, thereby addressing limitations imposed by sensors and imaging conditions. However, the inherent characteristics of remote sensing images, including diverse ground object types and complex details, pose significant challenges to achieving high-quality reconstruction. Existing methods typically employ a uniform structure to process various types of ground objects without distinction, making it difficult to adapt to the complex characteristics of remote sensing images. To address this issue, we introduce a Mixture of Experts (MoE) model and design a set of heterogeneous experts. These experts are organized into multiple expert groups, where experts within each group are homogeneous while being heterogeneous across groups. This design ensures that specialized activation parameters can be employed to handle the diverse and intricate details of ground objects effectively. To better accommodate the heterogeneous experts, we propose a multi-level feature aggregation strategy to guide the routing process. Additionally, we develop a dual-routing mechanism to adaptively select the optimal expert for each pixel. Experiments conducted on the UCMerced and AID datasets demonstrate that our proposed method achieves superior SR reconstruction accuracy compared to state-of-the-art methods. The code will be available at https://github.com/Mr-Bamboo/MFG-HMoE.

Test-Time Scaling (TTS) is an important method for improving the performance of Large Language Models (LLMs) by using additional computation during the inference phase. However, current studies do not systematically analyze how policy models, Process Reward Models (PRMs), and problem difficulty influence TTS. This lack of analysis limits the understanding and practical use of TTS methods. In this paper, we focus on two core questions: (1) What is the optimal approach to scale test-time computation across different policy models, PRMs, and problem difficulty levels? (2) To what extent can extended computation improve the performance of LLMs on complex tasks, and can smaller language models outperform larger ones through this approach? Through comprehensive experiments on MATH-500 and challenging AIME24 tasks, we have the following observations: (1) The compute-optimal TTS strategy is highly dependent on the choice of policy model, PRM, and problem difficulty. (2) With our compute-optimal TTS strategy, extremely small policy models can outperform larger models. For example, a 1B LLM can exceed a 405B LLM on MATH-500. Moreover, on both MATH-500 and AIME24, a 0.5B LLM outperforms GPT-4o, a 3B LLM surpasses a 405B LLM, and a 7B LLM beats o1 and DeepSeek-R1, while with higher inference efficiency. These findings show the significance of adapting TTS strategies to the specific characteristics of each task and model and indicate that TTS is a promising approach for enhancing the reasoning abilities of LLMs.

First-order logic (FOL) reasoning, which involves sequential deduction, is pivotal for intelligent systems and serves as a valuable task for evaluating reasoning capabilities, particularly in chain-of-thought (CoT) contexts. Existing benchmarks often rely on extensive human annotation or handcrafted templates, making it difficult to achieve the necessary complexity, scalability, and diversity for robust evaluation. To address these limitations, we propose a novel framework called ProverGen that synergizes the generative strengths of Large Language Models (LLMs) with the rigor and precision of symbolic provers, enabling the creation of a scalable, diverse, and high-quality FOL reasoning dataset, ProverQA. ProverQA is also distinguished by its inclusion of accessible and logically coherent intermediate reasoning steps for each problem. Our evaluation shows that state-of-the-art LLMs struggle to solve ProverQA problems, even with CoT prompting, highlighting the dataset's challenging nature. We also finetune Llama3.1-8B-Instruct on a separate training set generated by our framework. The finetuned model demonstrates consistent improvements on both in-distribution and out-of-distribution test sets, suggesting the value of our proposed data generation framework. Code available at: https://github.com/opendatalab/ProverGen, Comment: Accepted by ICLR 2025

Many-body electron-hole interactions are essential for understanding non-linear optical processes and ultrafast spectroscopy of materials. Recent first principles approaches based on nonequilibrium Green's function formalisms, such as the time-dependent adiabatic GW (TD-aGW) approach, can predict the nonequilibrium dynamics of excited states including electron-hole interactions. However, the high dimensionality of the electron-hole kernel poses significant computational challenges for scalability. Here, we develop a data-driven low-rank approximation for the electron-hole kernel, leveraging localized excitonic effects in the Hilbert space of crystalline systems. Through singular value decomposition (SVD) analysis, we show that the subspace of non-zero singular values, containing the key information of the electron-hole kernel, retains a small size even as the k-grid grows, ensuring computational feasibility with extremely dense k-grids for converged calculations. Utilizing this low-rank property, we achieve at least 95% compression of the kernel and an order-of-magnitude speedup of TD-aGW calculations. Our method, rooted in physical interpretability, outperforms existing machine learning approaches by avoiding intensive training processes and eliminating time-accumulated errors, providing a general framework for high-throughput, nonequilibrium simulation of light-driven dynamics in materials., Comment: 14 pages, 8 figures

The quantification of audio aesthetics remains a complex challenge in audio processing, primarily due to its subjective nature, which is influenced by human perception and cultural context. Traditional methods often depend on human listeners for evaluation, leading to inconsistencies and high resource demands. This paper addresses the growing need for automated systems capable of predicting audio aesthetics without human intervention. Such systems are crucial for applications like data filtering, pseudo-labeling large datasets, and evaluating generative audio models, especially as these models become more sophisticated. In this work, we introduce a novel approach to audio aesthetic evaluation by proposing new annotation guidelines that decompose human listening perspectives into four distinct axes. We develop and train no-reference, per-item prediction models that offer a more nuanced assessment of audio quality. Our models are evaluated against human mean opinion scores (MOS) and existing methods, demonstrating comparable or superior performance. This research not only advances the field of audio aesthetics but also provides open-source models and datasets to facilitate future work and benchmarking. We release our code and pre-trained model at: https://github.com/facebookresearch/audiobox-aesthetics, Comment: Repository: https://github.com/facebookresearch/audiobox-aesthetics Website: https://ai.meta.com/research/publications/meta-audiobox-aesthetics-unified-automatic-quality-assessment-for-speech-music-and-sound/

We report observations of solar wind turbulence derived from measurements by the Parker Solar Probe. Our findings reveal the emergence of finite magnetic helicity within the transition range of the turbulence, aligning with signatures of kinetic Alfv\'en waves (KAWs). Notably, as the wave scale transitions from super-ion to sub-ion scales, the ratio of KAWs with opposing signs of magnetic helicity initially increases from approximately 1 to 6 before returning to 1. This observation provides, for the first time, compelling evidence for the transition from imbalanced kinetic Alfv\'enic turbulence to balanced kinetic Alfv\'enic turbulence., Comment: 7 pages, 3 figures

Diffusion models have emerged as powerful tools for generative tasks, producing high-quality outputs across diverse domains. However, how the generated data responds to the initial noise perturbation in diffusion models remains under-explored, which hinders understanding the controllability of the sampling process. In this work, we first observe an interesting phenomenon: the relationship between the change of generation outputs and the scale of initial noise perturbation is highly linear through the diffusion ODE sampling. Then we provide both theoretical and empirical study to justify this linearity property of this input-output (noise-generation data) relationship. Inspired by these new insights, we propose a novel Controllable and Constrained Sampling method (CCS) together with a new controller algorithm for diffusion models to sample with desired statistical properties while preserving good sample quality. We perform extensive experiments to compare our proposed sampling approach with other methods on both sampling controllability and sampled data quality. Results show that our CCS method achieves more precisely controlled sampling while maintaining superior sample quality and diversity.

Whispering-gallery-mode (WGM) microcavities have emerged as a promising alternative to traditional ultrasound probes, offering high sensitivity and wide bandwidth. In our research, we propose a novel silica WGM microprobe device, with impressive Q factors up to 10^7.The side-coupled approach and special encapsulation design make the device small, robust, and capable of utilizing in both gaseous and liquid environments.We have successfully conducted photoacoustic (PA) imaging on various samples using this device which demonstrates a high sensitivity of 5.4 mPa/sqrt(Hz) and a board bandwidth of 41 MHz at -6 dB for ultrasound. What's more, it's capable of capturing the vibration spectrum of microparticles up to a few hundred megahertz. Our compact and lightweight device exhibits significant application potential in PA endoscopic detection, near-field ultrasound sensing and other aspects.

We develop a phenomenological model of strong imbalanced magnetohydrodynamic (MHD) turbulence that accounts for intermittency and the reflection of Alfven waves by spatial variations in the Alfven speed. Our model predicts the slopes of the inertial-range Elsasser power spectra, the scaling exponents of the higher-order Elsasser structure functions, and the way in which the parallel (to the magnetic field) length scale of the fluctuations varies with the perpendicular length scale. These predictions agree reasonably well with measurements of solar-wind turbulence from the Parker Solar Probe (PSP). In contrast to previous models of intermittency in balanced MHD turbulence, we find that intermittency in reflection-driven MHD turbulence increases the parallel wave numbers of the energetically dominant fluctuations at small perpendicular length scales. This, like the PSP measurements with which our model agrees, suggests that turbulence in the solar wind and solar corona may lead to more ion cyclotron heating than previously realized., Comment: 16 pages, 3 figures, accepted for publication in the Journal of Plasma Physics

Generative Adversarial Networks (GAN) have greatly influenced the development of computer vision and artificial intelligence in the past decade and also connected art and machine intelligence together. This book begins with a detailed introduction to the fundamental principles and historical development of GANs, contrasting them with traditional generative models and elucidating the core adversarial mechanisms through illustrative Python examples. The text systematically addresses the mathematical and theoretical underpinnings including probability theory, statistics, and game theory providing a solid framework for understanding the objectives, loss functions, and optimisation challenges inherent to GAN training. Subsequent chapters review classic variants such as Conditional GANs, DCGANs, InfoGAN, and LAPGAN before progressing to advanced training methodologies like Wasserstein GANs, GANs with gradient penalty, least squares GANs, and spectral normalisation techniques. The book further examines architectural enhancements and task-specific adaptations in generators and discriminators, showcasing practical implementations in high resolution image generation, artistic style transfer, video synthesis, text to image generation and other multimedia applications. The concluding sections offer insights into emerging research trends, including self-attention mechanisms, transformer-based generative models, and a comparative analysis with diffusion models, thus charting promising directions for future developments in both academic and applied settings.

The advent of next-generation video generation models like \textit{Sora} poses challenges for AI-generated content (AIGC) video quality assessment (VQA). These models substantially mitigate flickering artifacts prevalent in prior models, enable longer and complex text prompts and generate longer videos with intricate, diverse motion patterns. Conventional VQA methods designed for simple text and basic motion patterns struggle to evaluate these content-rich videos. To this end, we propose \textbf{CRAVE} (\underline{C}ontent-\underline{R}ich \underline{A}IGC \underline{V}ideo \underline{E}valuator), specifically for the evaluation of Sora-era AIGC videos. CRAVE proposes the multi-granularity text-temporal fusion that aligns long-form complex textual semantics with video dynamics. Additionally, CRAVE leverages the hybrid motion-fidelity modeling to assess temporal artifacts. Furthermore, given the straightforward prompts and content in current AIGC VQA datasets, we introduce \textbf{CRAVE-DB}, a benchmark featuring content-rich videos from next-generation models paired with elaborate prompts. Extensive experiments have shown that the proposed CRAVE achieves excellent results on multiple AIGC VQA benchmarks, demonstrating a high degree of alignment with human perception. All data and code will be publicly available at https://github.com/littlespray/CRAVE.

Large Language Models (LLMs) have revolutionized artificial intelligence with capabilities in reasoning, coding, and communication, driving innovation across industries. Their true potential depends on effective alignment to ensure correct, trustworthy and ethical behavior, addressing challenges like misinformation, hallucinations, bias and misuse. While existing Reinforcement Learning (RL)-based alignment methods are notoriously complex, direct optimization approaches offer a simpler alternative. In this work, we introduce a novel direct optimization approach for LLM alignment by drawing on established Information Retrieval (IR) principles. We present a systematic framework that bridges LLM alignment and IR methodologies, mapping LLM generation and reward models to IR's retriever-reranker paradigm. Building on this foundation, we propose LLM Alignment as Retriever Preference Optimization (LarPO), a new alignment method that enhances overall alignment quality. Extensive experiments validate LarPO's effectiveness with 38.9 % and 13.7 % averaged improvement on AlpacaEval2 and MixEval-Hard respectively. Our work opens new avenues for advancing LLM alignment by integrating IR foundations, offering a promising direction for future research., Comment: 26 pages

Exploratory Data Analysis (EDA) is a routine task for data analysts, often conducted using flexible computational notebooks. During EDA, data workers process, visualize, and interpret data tables, making decisions about subsequent analysis. However, the cell-by-cell programming approach, while flexible, can lead to disorganized code, making it difficult to trace the state of data tables across cells and increasing the cognitive load on data workers. This paper introduces NoteFlow, a notebook library that recommends charts as ``sight glasses'' for data tables, allowing users to monitor their dynamic updates throughout the EDA process. To ensure visual consistency and effectiveness, NoteFlow adapts chart encodings in response to data transformations, maintaining a coherent and insightful representation of the data. The proposed method was evaluated through user studies, demonstrating its ability to provide an overview of the EDA process and convey critical insights in the data tables.

Long-form generation is crucial for academic writing papers and repo-level code generation. Despite this, current models, including GPT-4o, still exhibit unsatisfactory performance. Existing methods that utilize preference learning with outcome supervision often fail to provide detailed feedback for extended contexts. This shortcoming can lead to content that does not fully satisfy query requirements, resulting in issues like length deviations, and diminished quality. In this paper, we propose enhancing long-form generation by incorporating process supervision. We employ Monte Carlo Tree Search to gather stepwise preference pairs, utilizing a global memory pool to maintain consistency. To address the issue of suboptimal candidate selection, we integrate external critiques to refine and improve the quality of the preference pairs. Finally, we apply step-level DPO using the collected stepwise preference pairs. Experimental results show that our method improves length and quality on long-form generation benchmarks, with almost lossless performance on general benchmarks across various model backbones.

Dense process rewards have proven a more effective alternative to the sparse outcome-level rewards in the inference-time scaling of large language models (LLMs), particularly in tasks requiring complex multi-step reasoning. While dense rewards also offer an appealing choice for the reinforcement learning (RL) of LLMs since their fine-grained rewards have the potential to address some inherent issues of outcome rewards, such as training efficiency and credit assignment, this potential remains largely unrealized. This can be primarily attributed to the challenges of training process reward models (PRMs) online, where collecting high-quality process labels is prohibitively expensive, making them particularly vulnerable to reward hacking. To address these challenges, we propose PRIME (Process Reinforcement through IMplicit rEwards), which enables online PRM updates using only policy rollouts and outcome labels through implict process rewards. PRIME combines well with various advantage functions and forgoes the dedicated reward model training phrase that existing approaches require, substantially reducing the development overhead. We demonstrate PRIME's effectiveness on competitional math and coding. Starting from Qwen2.5-Math-7B-Base, PRIME achieves a 15.1% average improvement across several key reasoning benchmarks over the SFT model. Notably, our resulting model, Eurus-2-7B-PRIME, surpasses Qwen2.5-Math-7B-Instruct on seven reasoning benchmarks with 10% of its training data., Comment: 20 pages. Model&Code&Data available at https://github.com/PRIME-RL/PRIME

Mixture-of-Experts (MoE) models are crucial for scaling model capacity while controlling inference costs. While integrating MoE into multimodal models like CLIP improves performance, training these models is notoriously challenging and expensive. We propose CLIP-Upcycling (CLIP-UP), an efficient alternative training strategy that converts a pre-trained dense CLIP model into a sparse MoE architecture. Through extensive experimentation with various settings and auxiliary losses, we demonstrate that CLIP-UP significantly reduces training complexity and cost. Remarkably, our sparse CLIP B/16 model, trained with CLIP-UP, outperforms its dense counterpart by 7.2% and 6.6% on COCO and Flickr30k text-to-image Recall@1 benchmarks respectively. It even surpasses the larger CLIP L/14 model on this task while using only 30% of the inference FLOPs. We further demonstrate the generalizability of our training recipe across different scales, establishing sparse upcycling as a practical and scalable approach for building efficient, high-performance CLIP models.

Vision-language navigation in unknown environments is crucial for mobile robots. In scenarios such as household assistance and rescue, mobile robots need to understand a human command, such as "find a person wearing black". We present a novel vision-language navigation (VL-Nav) system that integrates efficient spatial reasoning on low-power robots. Unlike prior methods that rely on a single image-level feature similarity to guide a robot, our method integrates pixel-wise vision-language features with curiosity-driven exploration. This approach enables robust navigation to human-instructed instances across diverse environments. We deploy VL-Nav on a four-wheel mobile robot and evaluate its performance through comprehensive navigation tasks in both indoor and outdoor environments, spanning different scales and semantic complexities. Remarkably, VL-Nav operates at a real-time frequency of 30 Hz with a Jetson Orin NX, highlighting its ability to conduct efficient vision-language navigation. Results show that VL-Nav achieves an overall success rate of 86.3%, outperforming previous methods by 44.15%.

Multi-task learning through composite loss functions is fundamental to modern deep learning, yet optimizing competing objectives remains challenging. We present new theoretical and practical approaches for addressing directional conflicts between loss terms, demonstrating their effectiveness in physics-informed neural networks (PINNs) where such conflicts are particularly challenging to resolve. Through theoretical analysis, we demonstrate how these conflicts limit first-order methods and show that second-order optimization naturally resolves them through implicit gradient alignment. We prove that SOAP, a recently proposed quasi-Newton method, efficiently approximates the Hessian preconditioner, enabling breakthrough performance in PINNs: state-of-the-art results on 10 challenging PDE benchmarks, including the first successful application to turbulent flows with Reynolds numbers up to 10,000, with 2-10x accuracy improvements over existing methods. We also introduce a novel gradient alignment score that generalizes cosine similarity to multiple gradients, providing a practical tool for analyzing optimization dynamics. Our findings establish frameworks for understanding and resolving gradient conflicts, with broad implications for optimization beyond scientific computing., Comment: 39 pages, 22 figures

We introduce MedXpertQA, a highly challenging and comprehensive benchmark to evaluate expert-level medical knowledge and advanced reasoning. MedXpertQA includes 4,460 questions spanning 17 specialties and 11 body systems. It includes two subsets, Text for text evaluation and MM for multimodal evaluation. Notably, MM introduces expert-level exam questions with diverse images and rich clinical information, including patient records and examination results, setting it apart from traditional medical multimodal benchmarks with simple QA pairs generated from image captions. MedXpertQA applies rigorous filtering and augmentation to address the insufficient difficulty of existing benchmarks like MedQA, and incorporates specialty board questions to improve clinical relevance and comprehensiveness. We perform data synthesis to mitigate data leakage risk and conduct multiple rounds of expert reviews to ensure accuracy and reliability. We evaluate 16 leading models on MedXpertQA. Moreover, medicine is deeply connected to real-world decision-making, providing a rich and representative setting for assessing reasoning abilities beyond mathematics and code. To this end, we develop a reasoning-oriented subset to facilitate the assessment of o1-like models.

Long-context large language models (LLMs) have achieved remarkable advancements, driven by techniques like Rotary Position Embedding (RoPE) (Su et al., 2023) and its extensions (Chen et al., 2023; Liu et al., 2024c; Peng et al., 2023). By adjusting RoPE parameters and incorporating training data with extended contexts, we can train performant models with considerably longer input sequences. However, existing RoPE-based methods exhibit performance limitations when applied to extended context lengths. This paper presents a comprehensive analysis of various attention mechanisms, including RoPE, No Positional Embedding (NoPE), and Query-Key Normalization (QK-Norm), identifying their strengths and shortcomings in long-context modeling. Our investigation identifies distinctive attention patterns in these methods and highlights their impact on long-context performance, providing valuable insights for architectural design. Building on these findings, we propose a novel architectural based on a hybrid attention mechanism that not only surpasses conventional RoPE-based transformer models in long context tasks but also achieves competitive performance on benchmarks requiring shorter context lengths.

Rationale and Objectives: Early prediction of pathological complete response (pCR) can facilitate personalized treatment for breast cancer patients. To improve prediction accuracy at the early time point of neoadjuvant chemotherapy, we proposed a two-stage dual-task learning strategy to train a deep neural network for early prediction of pCR using early-treatment magnetic resonance images. Methods: We developed and validated the two-stage dual-task learning strategy using the dataset from the national-wide, multi-institutional I-SPY2 clinical trial, which included dynamic contrast-enhanced magnetic resonance images acquired at three time points: pretreatment (T0), after 3 weeks (T1), and after 12 weeks of treatment (T2). First, we trained a convolutional long short-term memory network to predict pCR and extract the latent space image features at T2. At the second stage, we trained a dual-task network to simultaneously predict pCR and the image features at T2 using images from T0 and T1. This allowed us to predict pCR earlier without using images from T2. Results: The conventional single-stage single-task strategy gave an area under the receiver operating characteristic curve (AUROC) of 0.799 for pCR prediction using all the data at time points T0 and T1. By using the proposed two-stage dual-task learning strategy, the AUROC was improved to 0.820. Conclusions: The proposed two-stage dual-task learning strategy can improve model performance significantly (p=0.0025) for predicting pCR at the early stage (3rd week) of neoadjuvant chemotherapy. The early prediction model can potentially help physicians to intervene early and develop personalized plans at the early stage of chemotherapy.

Prohibited item detection based on X-ray images is one of the most effective security inspection methods. However, the foreground-background feature coupling caused by the overlapping phenomenon specific to X-ray images makes general detectors designed for natural images perform poorly. To address this issue, we propose a Category Semantic Prior Contrastive Learning (CSPCL) mechanism, which aligns the class prototypes perceived by the classifier with the content queries to correct and supplement the missing semantic information responsible for classification, thereby enhancing the model sensitivity to foreground features.To achieve this alignment, we design a specific contrastive loss, CSP loss, which includes Intra-Class Truncated Attraction (ITA) loss and Inter-Class Adaptive Repulsion (IAR) loss, and outperforms classic N-pair loss and InfoNCE loss. Specifically, ITA loss leverages class prototypes to attract intra-class category-specific content queries while preserving necessary distinctiveness. IAR loss utilizes class prototypes to adaptively repel inter-class category-specific content queries based on the similarity between class prototypes, helping disentangle features of similar categories.CSPCL is general and can be easily integrated into Deformable DETR-based models. Extensive experiments on the PIXray and OPIXray datasets demonstrate that CSPCL significantly enhances the performance of various state-of-the-art models without increasing complexity.The code will be open source once the paper is accepted., Comment: 10 pages

Assisted Reality (aR) is a subfield of Augmented Reality (AR) that overlays information onto a user's immediate view via see-through head-mounted displays (OST-HMDs). This technology has proven to be effective and energy-efficient to support the user and information interaction for everyday wearable intelligent systems. The aR viewing experience, however, is affected by varying real-world backgrounds, lighting, and user movements, which makes designing for aR challenging. Designers have to test their designs in-situ across multiple real-world settings, which can be time-consuming and labor-intensive. We propose SimulataR, a cost-effective desktop-based approach for rapid aR prototyping using first-person-view context videos blended with design prototypes to simulate an aR experience. A field study involving 12 AR users comparing SimulataR to real OST-HMDs found that SimulataR can approximate the aR experience, particularly for indoors and in low-to-moderate lit outdoor environments. Case studies with two designers who used SimulataR in their design process demonstrates the potential of design-blended videos for rapid aR prototyping.

We present a novel flow-based kinetic approach, inspired by continuous normalizing flows, for plasma simulation that unifies the complementary strengths of direct Vlasov solvers and particle-based methods. By tracking the distribution function along the characteristic curves defined by the Newton--Lorentz equations, our method directly computes f(z(t)) at selected points in phase space without reliance on Monte Carlo sampling. We employ a scatter-point integration scheme using smoothing kernels reminiscent of Smoothed Particle Hydrodynamics (SPH), to calculate field quantities and moments, achieving higher accuracy with far fewer markers compared to Particle-in-Cell (PIC) methods. Unlike PIC, our approach supports strategic marker placement and dynamic refinement in regions of interest, thus reducing sampling noise and computational overhead. This capability is particularly advantageous in high-density plasmas, where PIC's particle requirements can be prohibitive. In addition, the method naturally accommodates collisional effects via an augmented phase-space flow description ensuring robust handling of both collisionless and collisional plasmas. Our simulations of Landau damping, two-stream instability, and collisional relaxation demonstrate reduced noise, accurate phase-space resolution with significantly fewer markers, and robust energy conservation. Moreover, the independent characteristic curves and local scatter integration are highly amenable to GPU acceleration, enabling efficient large-scale simulations. Overall, this flow-based framework offers a powerful, flexible, and computationally efficient alternative to traditional particle methods for kinetic plasma dynamics, with potential applications spanning inertial confinement, Zpinch, and other complex kinetic systems., Comment: 49 pages, 6 figures

Understanding the chemical structure from a graphical representation of a molecule is a challenging image caption task that would greatly benefit molecule-centric scientific discovery. Variations in molecular images and caption subtasks pose a significant challenge in both image representation learning and task modeling. Yet, existing methods only focus on a specific caption task that translates a molecular image into its graph structure, i.e., OCSR. In this paper, we propose the Optical Chemical Structure Understanding (OCSU) task, which extends OCSR to molecular image caption from motif level to molecule level and abstract level. We present two approaches for that, including an OCSR-based method and an end-to-end OCSR-free method. The proposed Double-Check achieves SOTA OCSR performance on real-world patent and journal article scenarios via attentive feature enhancement for local ambiguous atoms. Cascading with SMILES-based molecule understanding methods, it can leverage the power of existing task-specific models for OCSU. While Mol-VL is an end-to-end optimized VLM-based model. An OCSU dataset, Vis-CheBI20, is built based on the widely used CheBI20 dataset for training and evaluation. Extensive experimental results on Vis-CheBI20 demonstrate the effectiveness of the proposed approaches. Improving OCSR capability can lead to a better OCSU performance for OCSR-based approach, and the SOTA performance of Mol-VL demonstrates the great potential of end-to-end approach.

We introduce Qwen2.5-1M, a series of models that extend the context length to 1 million tokens. Compared to the previous 128K version, the Qwen2.5-1M series have significantly enhanced long-context capabilities through long-context pre-training and post-training. Key techniques such as long data synthesis, progressive pre-training, and multi-stage supervised fine-tuning are employed to effectively enhance long-context performance while reducing training costs. To promote the use of long-context models among a broader user base, we present and open-source our inference framework. This framework includes a length extrapolation method that can expand the model context lengths by at least four times, or even more, without additional training. To reduce inference costs, we implement a sparse attention method along with chunked prefill optimization for deployment scenarios and a sparsity refinement method to improve precision. Additionally, we detail our optimizations in the inference engine, including kernel optimization, pipeline parallelism, and scheduling optimization, which significantly enhance overall inference performance. By leveraging our inference framework, the Qwen2.5-1M models achieve a remarkable 3x to 7x prefill speedup in scenarios with 1 million tokens of context. This framework provides an efficient and powerful solution for developing applications that require long-context processing using open-source models. The Qwen2.5-1M series currently includes the open-source models Qwen2.5-7B-Instruct-1M and Qwen2.5-14B-Instruct-1M, as well as the API-accessed model Qwen2.5-Turbo. Evaluations show that Qwen2.5-1M models have been greatly improved in long-context tasks without compromising performance in short-context scenarios. Specifically, the Qwen2.5-14B-Instruct-1M model significantly outperforms GPT-4o-mini in long-context tasks and supports contexts eight times longer.

Evolutionary Reinforcement Learning (EvoRL) has emerged as a promising approach to overcoming the limitations of traditional reinforcement learning (RL) by integrating the Evolutionary Computation (EC) paradigm with RL. However, the population-based nature of EC significantly increases computational costs, thereby restricting the exploration of algorithmic design choices and scalability in large-scale settings. To address this challenge, we introduce $\texttt{$\textbf{EvoRL}$}$, the first end-to-end EvoRL framework optimized for GPU acceleration. The framework executes the entire training pipeline on accelerators, including environment simulations and EC processes, leveraging hierarchical parallelism through vectorization and compilation techniques to achieve superior speed and scalability. This design enables the efficient training of large populations on a single machine. In addition to its performance-oriented design, $\texttt{$\textbf{EvoRL}$}$ offers a comprehensive platform for EvoRL research, encompassing implementations of traditional RL algorithms (e.g., A2C, PPO, DDPG, TD3, SAC), Evolutionary Algorithms (e.g., CMA-ES, OpenES, ARS), and hybrid EvoRL paradigms such as Evolutionary-guided RL (e.g., ERL, CEM-RL) and Population-Based AutoRL (e.g., PBT). The framework's modular architecture and user-friendly interface allow researchers to seamlessly integrate new components, customize algorithms, and conduct fair benchmarking and ablation studies. The project is open-source and available at: https://github.com/EMI-Group/evorl.