Search

While recent advancements in aligning Large Language Models (LLMs) with recommendation tasks have shown great potential and promising performance overall, these aligned recommendation LLMs still face challenges in complex scenarios. This is primarily due to the current alignment approach focusing on optimizing LLMs to generate user feedback directly, without incorporating deliberation. To overcome this limitation and develop more reliable LLMs for recommendations, we propose a new Deliberative Recommendation task, which incorporates explicit reasoning about user preferences as an additional alignment goal. We then introduce the Reasoning-powered Recommender framework for deliberative user preference alignment, designed to enhance reasoning capabilities by utilizing verbalized user feedback in a step-wise manner to tackle this task. The framework employs collaborative step-wise experts and tailored training strategies for each expert. Experimental results across three real-world datasets demonstrate the rationality of the deliberative task formulation and the superior performance of the proposed framework in improving both prediction accuracy and reasoning quality.

Recent advancements have highlighted the Mamba framework, a state-space model known for its efficiency in capturing long-range dependencies with linear computational complexity. While Mamba has shown competitive performance in medical image segmentation, it encounters difficulties in modeling local features due to the sporadic nature of traditional location-based scanning methods and the complex, ambiguous boundaries often present in medical images. To overcome these challenges, we propose Uncertainty-Driven Mamba (UD-Mamba), which redefines the pixel-order scanning process by incorporating channel uncertainty into the scanning mechanism. UD-Mamba introduces two key scanning techniques: 1) sequential scanning, which prioritizes regions with high uncertainty by scanning in a row-by-row fashion, and 2) skip scanning, which processes columns vertically, moving from high-to-low or low-to-high uncertainty at fixed intervals. Sequential scanning efficiently clusters high-uncertainty regions, such as boundaries and foreground objects, to improve segmentation precision, while skip scanning enhances the interaction between background and foreground regions, allowing for timely integration of background information to support more accurate foreground inference. Recognizing the advantages of scanning from certain to uncertain areas, we introduce four learnable parameters to balance the importance of features extracted from different scanning methods. Additionally, a cosine consistency loss is employed to mitigate the drawbacks of transitioning between uncertain and certain regions during the scanning process. Our method demonstrates robust segmentation performance, validated across three distinct medical imaging datasets involving pathology, dermatological lesions, and cardiac tasks., Comment: 19 pages

Recent advances in CV and NLP have inspired researchers to develop general-purpose graph foundation models through pre-training across diverse domains. However, a fundamental challenge arises from the substantial differences in graph topologies across domains. Additionally, real-world graphs are often sparse and prone to noisy connections and adversarial attacks. To address these issues, we propose the Multi-Domain Graph Foundation Model (MDGFM), a unified framework that aligns and leverages cross-domain topological information to facilitate robust knowledge transfer. MDGFM bridges different domains by adaptively balancing features and topology while refining original graphs to eliminate noise and align topological structures. To further enhance knowledge transfer, we introduce an efficient prompt-tuning approach. By aligning topologies, MDGFM not only improves multi-domain pre-training but also enables robust knowledge transfer to unseen domains. Theoretical analyses provide guarantees of MDGFM's effectiveness and domain generalization capabilities. Extensive experiments on both homophilic and heterophilic graph datasets validate the robustness and efficacy of our method.

We introduce OmniRL, a highly generalizable in-context reinforcement learning (ICRL) model that is meta-trained on hundreds of thousands of diverse tasks. These tasks are procedurally generated by randomizing state transitions and rewards within Markov Decision Processes. To facilitate this extensive meta-training, we propose two key innovations: 1. An efficient data synthesis pipeline for ICRL, which leverages the interaction histories of diverse behavior policies; and 2. A novel modeling framework that integrates both imitation learning and reinforcement learning (RL) within the context, by incorporating prior knowledge. For the first time, we demonstrate that in-context learning (ICL) alone, without any gradient-based fine-tuning, can successfully tackle unseen Gymnasium tasks through imitation learning, online RL, or offline RL. Additionally, we show that achieving generalized ICRL capabilities-unlike task identification-oriented few-shot learning-critically depends on long trajectories generated by variant tasks and diverse behavior policies. By emphasizing the potential of ICL and departing from pre-training focused on acquiring specific skills, we further underscore the significance of meta-training aimed at cultivating the ability of ICL itself., Comment: Preprint

The K-long and muon detector (KLM) constitutes the outer-most volume of the Belle II spectrometer at the interaction region of the SuperKEKB collider in Tsukuba, Japan. The KLM detector was partially upgraded since the Belle experiment by replacing many of its resistive-plate chambers with scintillators containing wavelength-shifting fibers and instrumenting it with silicon photomultipliers. We describe the readout electronics, firmware, and software created to control and acquire data from the scintillators and resistive-plate chambers., Comment: 14 pages, 20 figures. To be submitted to Nuclear Instruments and Methods in Physics Research Section A

Streaming multi-talker speech translation is a task that involves not only generating accurate and fluent translations with low latency but also recognizing when a speaker change occurs and what the speaker's gender is. Speaker change information can be used to create audio prompts for a zero-shot text-to-speech system, and gender can help to select speaker profiles in a conventional text-to-speech model. We propose to tackle streaming speaker change detection and gender classification by incorporating speaker embeddings into a transducer-based streaming end-to-end speech translation model. Our experiments demonstrate that the proposed methods can achieve high accuracy for both speaker change detection and gender classification.

3D articulated objects modeling has long been a challenging problem, since it requires to capture both accurate surface geometries and semantically meaningful and spatially precise structures, parts, and joints. Existing methods heavily depend on training data from a limited set of handcrafted articulated object categories (e.g., cabinets and drawers), which restricts their ability to model a wide range of articulated objects in an open-vocabulary context. To address these limitations, we propose Articulate Anymesh, an automated framework that is able to convert any rigid 3D mesh into its articulated counterpart in an open-vocabulary manner. Given a 3D mesh, our framework utilizes advanced Vision-Language Models and visual prompting techniques to extract semantic information, allowing for both the segmentation of object parts and the construction of functional joints. Our experiments show that Articulate Anymesh can generate large-scale, high-quality 3D articulated objects, including tools, toys, mechanical devices, and vehicles, significantly expanding the coverage of existing 3D articulated object datasets. Additionally, we show that these generated assets can facilitate the acquisition of new articulated object manipulation skills in simulation, which can then be transferred to a real robotic system. Our Github website is https://articulate-anymesh.github.io.

The advent of 6G communication demands seamlessly integrated terminals operating above 100 GHz with low power consumption for human-centric applications. In this work, we report high-performance, flexible radio-frequency (RF) transistors based on aligned carbon nanotube (CNT) arrays, achieving, for the first time, as-measured current gain cutoff frequency ($f_{\mathrm{T}}$) and power gain cutoff frequency ($f_{\mathrm{max}}$) both exceeding 100 GHz. Electro-thermal co-design improves both heat dissipation and RF performance, despite the low thermal conductivity of the flexible substrate. The transistors deliver 0.947 mA/ $\mathrm{\mu}$m on-state current and 0.728 mS/ $\mathrm{\mu}$m transconductance. Peak extrinsic $f_{\mathrm{T}}$ and $f_{\mathrm{max}}$ reach 152 GHz and 102 GHz, with low power consumption of 199 mW/mm and 147 mW/mm, respectively, setting new performance records for flexible CNT-based RF transistors by nearly 100$\times$, outperforming all other flexible RF devices. Additionally, flexible RF amplifiers achieve output power of 64 mW/mm and power gain of 11 dB in the K-band (18 GHz), marking a significant milestone in the development of flexible RF technologies for next-generation wireless communication systems., Comment: 34 pages, 15 figures

A smart ring is a wearable electronic device in the form of a ring that incorporates diverse sensors and computing technologies to perform a variety of functions. Designed for use with fingers, smart rings are capable of sensing more subtle and abundant hand movements, thus making them a good platform for interaction. Meanwhile, fingers are abundant with blood vessels and nerve endings and accustomed to wearing rings, providing an ideal site for continuous health monitoring through smart rings, which combine comfort with the ability to capture vital biometric data, making them suitable for all-day wear. We collected in total of 206 smart ring-related publications and conducted a systematic literature review. We provide a taxonomy regarding the sensing and feedback modalities, applications, and phenomena. We review and categorize these literatures into four main areas: (1) interaction - input, (2) interaction - output, (3) passive sensing - in body feature, (4) passive sensing - out body activity. This comprehensive review highlights the current advancements within the field of smart ring and identifies potential areas for future research.

3D visual grounding (3DVG) is challenging because of the requirement of understanding on visual information, language and spatial relationships. While supervised approaches have achieved superior performance, they are constrained by the scarcity and high cost of 3D vision-language datasets. On the other hand, LLM/VLM based agents are proposed for 3DVG, eliminating the need for training data. However, these methods incur prohibitive time and token costs during inference. To address the challenges, we introduce a novel training-free symbolic framework for 3D visual grounding, namely Evolvable Symbolic Visual Grounder, that offers significantly reduced inference costs compared to previous agent-based methods while maintaining comparable performance. EaSe uses LLM generated codes to compute on spatial relationships. EaSe also implements an automatic pipeline to evaluate and optimize the quality of these codes and integrate VLMs to assist in the grounding process. Experimental results demonstrate that EaSe achieves 52.9% accuracy on Nr3D dataset and 49.2% Acc@0.25 on ScanRefer, which is top-tier among training-free methods. Moreover, it substantially reduces the inference time and cost, offering a balanced trade-off between performance and efficiency. Codes are available at https://github.com/OpenRobotLab/EaSe.

Aligning visual features with language embeddings is a key challenge in vision-language models (VLMs). The performance of such models hinges on having a good connector that maps visual features generated by a vision encoder to a shared embedding space with the LLM while preserving semantic similarity. Existing connectors, such as multilayer perceptrons (MLPs), often produce out-of-distribution or noisy inputs, leading to misalignment between the modalities. In this work, we propose a novel vision-text alignment method, AlignVLM, that maps visual features to a weighted average of LLM text embeddings. Our approach leverages the linguistic priors encoded by the LLM to ensure that visual features are mapped to regions of the space that the LLM can effectively interpret. AlignVLM is particularly effective for document understanding tasks, where scanned document images must be accurately mapped to their textual content. Our extensive experiments show that AlignVLM achieves state-of-the-art performance compared to prior alignment methods. We provide further analysis demonstrating improved vision-text feature alignment and robustness to noise.

This paper presents a novel approach to Visual Inertial Odometry (VIO), focusing on the initialization and feature matching modules. Existing methods for initialization often suffer from either poor stability in visual Structure from Motion (SfM) or fragility in solving a huge number of parameters simultaneously. To address these challenges, we propose a new pipeline for visual inertial initialization that robustly handles various complex scenarios. By tightly coupling gyroscope measurements, we enhance the robustness and accuracy of visual SfM. Our method demonstrates stable performance even with only four image frames, yielding competitive results. In terms of feature matching, we introduce a hybrid method that combines optical flow and descriptor-based matching. By leveraging the robustness of continuous optical flow tracking and the accuracy of descriptor matching, our approach achieves efficient, accurate, and robust tracking results. Through evaluation on multiple benchmarks, our method demonstrates state-of-the-art performance in terms of accuracy and success rate. Additionally, a video demonstration on mobile devices showcases the practical applicability of our approach in the field of Augmented Reality/Virtual Reality (AR/VR).

Large language models (LLMs) have shown significant promise across various medical applications, with ophthalmology being a notable area of focus. Many ophthalmic tasks have shown substantial improvement through the integration of LLMs. However, before these models can be widely adopted in clinical practice, evaluating their capabilities and identifying their limitations is crucial. To address this research gap and support the real-world application of LLMs, we introduce the OphthBench, a specialized benchmark designed to assess LLM performance within the context of Chinese ophthalmic practices. This benchmark systematically divides a typical ophthalmic clinical workflow into five key scenarios: Education, Triage, Diagnosis, Treatment, and Prognosis. For each scenario, we developed multiple tasks featuring diverse question types, resulting in a comprehensive benchmark comprising 9 tasks and 591 questions. This comprehensive framework allows for a thorough assessment of LLMs' capabilities and provides insights into their practical application in Chinese ophthalmology. Using this benchmark, we conducted extensive experiments and analyzed the results from 39 popular LLMs. Our evaluation highlights the current gap between LLM development and its practical utility in clinical settings, providing a clear direction for future advancements. By bridging this gap, we aim to unlock the potential of LLMs and advance their development in ophthalmology.

Humanoid robots hold the potential for unparalleled versatility in performing human-like, whole-body skills. However, achieving agile and coordinated whole-body motions remains a significant challenge due to the dynamics mismatch between simulation and the real world. Existing approaches, such as system identification (SysID) and domain randomization (DR) methods, often rely on labor-intensive parameter tuning or result in overly conservative policies that sacrifice agility. In this paper, we present ASAP (Aligning Simulation and Real-World Physics), a two-stage framework designed to tackle the dynamics mismatch and enable agile humanoid whole-body skills. In the first stage, we pre-train motion tracking policies in simulation using retargeted human motion data. In the second stage, we deploy the policies in the real world and collect real-world data to train a delta (residual) action model that compensates for the dynamics mismatch. Then, ASAP fine-tunes pre-trained policies with the delta action model integrated into the simulator to align effectively with real-world dynamics. We evaluate ASAP across three transfer scenarios: IsaacGym to IsaacSim, IsaacGym to Genesis, and IsaacGym to the real-world Unitree G1 humanoid robot. Our approach significantly improves agility and whole-body coordination across various dynamic motions, reducing tracking error compared to SysID, DR, and delta dynamics learning baselines. ASAP enables highly agile motions that were previously difficult to achieve, demonstrating the potential of delta action learning in bridging simulation and real-world dynamics. These results suggest a promising sim-to-real direction for developing more expressive and agile humanoids., Comment: Project website: https://agile.human2humanoid.com/

Gravitational-wave (GW) ringdown signals from black holes (BHs) encode crucial information about the gravitational dynamics in the strong-field regime, which offers unique insights into BH properties. In the future, the improving sensitivity of GW detectors is to enable the extraction of multiple quasi-normal modes (QNMs) from ringdown signals. However, incorporating multiple modes drastically enlarges the parameter space, posing computational challenges to data analysis. Inspired by the $F$-statistic method in the continuous GW searches, we develope an algorithm, dubbed as FIREFLY, for accelerating the ringdown signal analysis. FIREFLY analytically marginalizes the amplitude and phase parameters of QNMs to reduce the computational cost and speed up the full-parameter inference from hours to minutes, while achieving consistent posterior and evidence. The acceleration becomes more significant when more QNMs are considered. Rigorously based on the principle of Bayesian inference and importance sampling, our method is statistically interpretable, flexible in prior choice, and compatible with various advanced sampling techniques, providing a new perspective for accelerating future GW data analysis., Comment: 21 pages, 14 figures

Contextual linear dueling bandits have recently garnered significant attention due to their widespread applications in important domains such as recommender systems and large language models. Classical dueling bandit algorithms are typically only applicable to a single agent. However, many applications of dueling bandits involve multiple agents who wish to collaborate for improved performance yet are unwilling to share their data. This motivates us to draw inspirations from federated learning, which involves multiple agents aiming to collaboratively train their neural networks via gradient descent (GD) without sharing their raw data. Previous works have developed federated linear bandit algorithms which rely on closed-form updates of the bandit parameters (e.g., the linear function parameter) to achieve collaboration. However, in linear dueling bandits, the linear function parameter lacks a closed-form expression and its estimation requires minimizing a loss function. This renders these previous methods inapplicable. In this work, we overcome this challenge through an innovative and principled combination of online gradient descent (for minimizing the loss function to estimate the linear function parameters) and federated learning, hence introducing the first federated linear dueling bandit algorithms. Through rigorous theoretical analysis, we prove that our algorithms enjoy a sub-linear upper bound on its cumulative regret. We also use empirical experiments to demonstrate the effectiveness of our algorithms and the practical benefit of collaboration.

Direct Preference Optimization (DPO) has shown effectiveness in aligning multi-modal large language models (MLLM) with human preferences. However, existing methods exhibit an imbalanced responsiveness to the data of varying hardness, tending to overfit on the easy-to-distinguish data while underfitting on the hard-to-distinguish data. In this paper, we propose Data- and Model-aware DPO (DAMA) to dynamically adjust the optimization process from two key aspects: (1) a data-aware strategy that incorporates data hardness, and (2) a model-aware strategy that integrates real-time model responses. By combining the two strategies, DAMA enables the model to effectively adapt to data with varying levels of hardness. Extensive experiments on five benchmarks demonstrate that DAMA not only significantly enhances the trustworthiness, but also improves the effectiveness over general tasks. For instance, on the Object-HalBench, our DAMA-7B reduces response-level and mentioned-level hallucination by 90.0% and 95.3%, respectively, surpassing the performance of GPT-4V.

Multi-agent reinforcement learning (MARL) has made significant progress, largely fueled by the development of specialized testbeds that enable systematic evaluation of algorithms in controlled yet challenging scenarios. However, existing testbeds often focus on purely virtual simulations or limited robot morphologies such as robotic arms, quadrupeds, and humanoids, leaving high-mobility platforms with real-world physical constraints like drones underexplored. To bridge this gap, we present VolleyBots, a new MARL testbed where multiple drones cooperate and compete in the sport of volleyball under physical dynamics. VolleyBots features a turn-based interaction model under volleyball rules, a hierarchical decision-making process that combines motion control and strategic play, and a high-fidelity simulation for seamless sim-to-real transfer. We provide a comprehensive suite of tasks ranging from single-drone drills to multi-drone cooperative and competitive tasks, accompanied by baseline evaluations of representative MARL and game-theoretic algorithms. Results in simulation show that while existing algorithms handle simple tasks effectively, they encounter difficulty in complex tasks that require both low-level control and high-level strategy. We further demonstrate zero-shot deployment of a simulation-learned policy to real-world drones, highlighting VolleyBots' potential to propel MARL research involving agile robotic platforms. The project page is at https://sites.google.com/view/thu-volleybots/home.

Dual active galactic nuclei (AGNs) with comparable masses are commonly witnessed among the major merged galaxies with interaction remnants. Considering almost every massive galaxy is associated with multiple dwarf satellites around it, minor mergers involving galaxies with disproportional stellar masses should be much more common than major mergers, which would naturally lead to black hole (BH) pairs with significantly different masses. However, dual AGNs generated by minor mergers involving one or two dwarf galaxies are exceptionally rare and understudied. Moreover, good estimates of the masses of both BHs are not yet available to test this idea. Here we report the evidence of a dual AGN candidate with mass ratio $\sim$7:1 located in an undisturbed disk galaxy. We identify the central BH with mass of $9.4 \times 10^6M_\odot$ from its radio emission as well as AGN-driven galactic-scale biconical outflows. The off-centered BH generates obvious broad and narrow emission-line regions, which gives us a robust estimation of a $1.3 \times 10^6M_\odot$ BH mass. We explore alternative scenarios for explaining the observational features of this system, including the complex gas kinematics triggered by central AGN activity and dust attenuation of the broad-line region of the central BH, finding that they failed to fully account for the kinematics of both the redshifted off-centered broad and narrow emission-line components., Comment: Published in ApJL on 30 January 2025; 14 pages, 7 figures, 1 table

Resting-state functional magnetic resonance imaging (rs-fMRI) and its derived functional connectivity networks (FCNs) have become critical for understanding neurological disorders. However, collaborative analyses and the generalizability of models still face significant challenges due to privacy regulations and the non-IID (non-independent and identically distributed) property of multiple data sources. To mitigate these difficulties, we propose Domain Adversarial Federated Learning (DAFed), a novel federated deep learning framework specifically designed for non-IID fMRI data analysis in multi-site settings. DAFed addresses these challenges through feature disentanglement, decomposing the latent feature space into domain-invariant and domain-specific components, to ensure robust global learning while preserving local data specificity. Furthermore, adversarial training facilitates effective knowledge transfer between labeled and unlabeled datasets, while a contrastive learning module enhances the global representation of domain-invariant features. We evaluated DAFed on the diagnosis of ASD and further validated its generalizability in the classification of AD, demonstrating its superior classification accuracy compared to state-of-the-art methods. Additionally, an enhanced Score-CAM module identifies key brain regions and functional connectivity significantly associated with ASD and MCI, respectively, uncovering shared neurobiological patterns across sites. These findings highlight the potential of DAFed to advance multi-site collaborative research in neuroimaging while protecting data confidentiality., Comment: 34pages, 13 figures

The Keck Planet Imager and Characterizer (KPIC) is a series of upgrades for the Keck II Adaptive Optics (AO) system and the NIRSPEC spectrograph to enable diffraction limited, high resolution (R>30000) spectroscopy of exoplanets and low mass companions in the K and L bands. Phase I consisted of single mode fiber injection/extraction units (FIU/FEU) used in conjunction with a H band pyramid wavefront sensor. The use of single mode fibers provides a gain in stellar rejection, a substantial reduction in sky background, and an extremely stable line spread function in the spectrograph. Phase II, deployed and commissioned in 2022, brought a 1000 actuator deformable mirror, beam shaping optics, a vortex mask, and other upgrades to the FIU/FEU. An additional service mission in 2024 extended operations down to y band, delivered an atmospheric dispersion corrector, and provided access to two laser frequency combs. KPIC phase II brings higher planet throughput, lower stellar leakage and many new observing modes which extend its ability to characterize exoplanets at high spectral resolution, building on the success of phase I. In this paper we present a description of the final phase II version of KPIC, along with results of system level laboratory testing and characterization showing the instrument's phase II throughput, stability, repeatability, and other key performance metrics prior to delivery and during installation at Keck. We outlined the capabilities of the various observing modes enabled by the new modules as well as efforts to compensate for static aberrations and non common path errors at Keck, which were issues that plagued phase I. Finally, we show results from commissioning., Comment: 36 pages, 16 figures

Information design (ID) explores how a sender influence the optimal behavior of receivers to achieve specific objectives. While ID originates from everyday human communication, existing game-theoretic and machine learning methods often model information structures as numbers, which limits many applications to toy games. This work leverages LLMs and proposes a verbalized framework in Bayesian persuasion (BP), which extends classic BP to real-world games involving human dialogues for the first time. Specifically, we map the BP to a verbalized mediator-augmented extensive-form game, where LLMs instantiate the sender and receiver. To efficiently solve the verbalized game, we propose a generalized equilibrium-finding algorithm combining LLM and game solver. The algorithm is reinforced with techniques including verbalized commitment assumptions, verbalized obedience constraints, and information obfuscation. Numerical experiments in dialogue scenarios, such as recommendation letters, courtroom interactions, and law enforcement, validate that our framework can both reproduce theoretical results in classic BP and discover effective persuasion strategies in more complex natural language and multi-stage scenarios., Comment: 63 pages, 21 figures

The discovery of superconductivity with onset temperature of ~80 K in pressurized bilayer Ruddlesden-Popper La3Ni2O7-d has attracted much attention. Despite intense research, determination of the exact oxygen content and understanding of the relationship between superconductivity and oxygen content remain a big challenge. Here, we report a systematical study on the structure and physical properties of La3Ni2O7-d polycrystalline powders which were prepared using sol-gel method at ambient pressure and then annealed under various oxygen pressure. We found that high pO2 annealing with slow cooling results in a new phase, which can be modeled using the hybrid single-layer-trilayer La3Ni2O7 or the tetragonal bilayer La3Ni2O7. Scanning transmission electron microscopy (STEM) measurements revealed significant single layers and trilayers after high oxygen pressure annealing, evidencing damage of the bilayer structure. The superconducting transition under high pressure became weak for high pO2 annealed samples, which is consistent with the damage of the bilayer structure. Our results reveal that the bilayer structure is fragile and post-annealing under near atmosphere pressure of oxygen is suitable to maintain bilayer structure and increase oxygen content at the same time., Comment: 5 figures and 1 table

The Hong-Ou-Mandel effect is a paradigmatic quantum phenomenon demonstrating the interference of two indistinguishable photons that are linearly coupled at a 50:50 beam splitter. Here, we transpose such a two-particle quantum interference effect to the nonlinear regime, when two single photons are impinging on a parametric down-conversion crystal. Formally, this transposition amounts to exchanging space and time variables, giving rise to an unknown form of timelike quantum interference. The two-photon component of the output state is a superposition of the incident photons being either transmitted or reborn, that is, replaced by indistinguishable substitutes due to their interaction with the nonlinear crystal. We experimentally demonstrate the suppression of the probability of detecting precisely one photon pair when the amplification gain is tuned to 2, which arises from the destructive interference between the transmitted and reborn photon pairs. This heretofore unobserved quantum manifestation of indistinguishability in time pushes nonlinear quantum interference towards a new regime with multiple photons. Hence, composing this effect with larger linear optical circuits should provide a tool to generate multimode quantum non-Gaussian states, which are essential resources for photonic quantum computers.

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

The rapid advancement of large models, driven by their exceptional abilities in learning and generalization through large-scale pre-training, has reshaped the landscape of Artificial Intelligence (AI). These models are now foundational to a wide range of applications, including conversational AI, recommendation systems, autonomous driving, content generation, medical diagnostics, and scientific discovery. However, their widespread deployment also exposes them to significant safety risks, raising concerns about robustness, reliability, and ethical implications. This survey provides a systematic review of current safety research on large models, covering Vision Foundation Models (VFMs), Large Language Models (LLMs), Vision-Language Pre-training (VLP) models, Vision-Language Models (VLMs), Diffusion Models (DMs), and large-model-based Agents. Our contributions are summarized as follows: (1) We present a comprehensive taxonomy of safety threats to these models, including adversarial attacks, data poisoning, backdoor attacks, jailbreak and prompt injection attacks, energy-latency attacks, data and model extraction attacks, and emerging agent-specific threats. (2) We review defense strategies proposed for each type of attacks if available and summarize the commonly used datasets and benchmarks for safety research. (3) Building on this, we identify and discuss the open challenges in large model safety, emphasizing the need for comprehensive safety evaluations, scalable and effective defense mechanisms, and sustainable data practices. More importantly, we highlight the necessity of collective efforts from the research community and international collaboration. Our work can serve as a useful reference for researchers and practitioners, fostering the ongoing development of comprehensive defense systems and platforms to safeguard AI models., Comment: 47 pages, 3 figures, 11 tables GitHub: https://github.com/xingjunm/Awesome-Large-Model-Safety

In this paper, we present WonderHuman to reconstruct dynamic human avatars from a monocular video for high-fidelity novel view synthesis. Previous dynamic human avatar reconstruction methods typically require the input video to have full coverage of the observed human body. However, in daily practice, one typically has access to limited viewpoints, such as monocular front-view videos, making it a cumbersome task for previous methods to reconstruct the unseen parts of the human avatar. To tackle the issue, we present WonderHuman, which leverages 2D generative diffusion model priors to achieve high-quality, photorealistic reconstructions of dynamic human avatars from monocular videos, including accurate rendering of unseen body parts. Our approach introduces a Dual-Space Optimization technique, applying Score Distillation Sampling (SDS) in both canonical and observation spaces to ensure visual consistency and enhance realism in dynamic human reconstruction. Additionally, we present a View Selection strategy and Pose Feature Injection to enforce the consistency between SDS predictions and observed data, ensuring pose-dependent effects and higher fidelity in the reconstructed avatar. In the experiments, our method achieves SOTA performance in producing photorealistic renderings from the given monocular video, particularly for those challenging unseen parts. The project page and source code can be found at https://wyiguanw.github.io/WonderHuman/.

Rydberg blockade effect provides a convenient platform for simulating locally constrained many-body systems, such as quantum dimer models and quantum loop models, especially their novel phases like topological orders and gapless quantum spin ice (QSI) phases. To discuss the possible phase diagram containing different QSIs in 3D Rydberg arrays, here, we have constructed an extended Rokhsar-Kivelson (RK) Hamiltonian with equal-weight-superposition ground state in different fillings at the RK point. Therefore, both the perfect QSIs with fixed local dimer filling and their monomer-doped states can be simulated directly by Monte Carlo sampling. Using single mode approximation, the excitations of dimers and monomers have also been explored in different fillings. We find that, in the thermodynamical limit, even doping a small amount of monomers can disrupt the topological structure and lead to the existence of off-diagonal long-range order. However, in a finite size (as in cold-atom experiment), the property of QSI will be kept in a certain region like a crossover after doping. The phase diagram containing different QSIs and off-diagonal order phases is proposed.

Selecting high-quality data can significantly improve the pretraining efficiency of large language models (LLMs). Existing methods generally rely on heuristic techniques and single-quality signals, limiting their ability to evaluate data quality comprehensively. In this work, we propose FIRE, a flexible and scalable framework for integrating multiple data quality raters, which allows for a comprehensive assessment of data quality across various dimensions. FIRE aligns multiple quality signals into a unified space, and integrates diverse data quality raters to provide a comprehensive quality signal for each data point. Further, we introduce a progressive data selection scheme based on FIRE that iteratively refines the selection of high-quality data points. Experiments on the SlimPajama dataset reveal that FIRE outperforms other data selection methods and significantly enhances the pretrained model across a wide range of downstream tasks, with a 2.9% average performance improvement over Random and reducing the FLOPs necessary to achieve a certain performance level by more than half., Comment: 19 pages, 11 figures

fMRI (functional Magnetic Resonance Imaging) visual decoding involves decoding the original image from brain signals elicited by visual stimuli. This often relies on manually labeled ROIs (Regions of Interest) to select brain voxels. However, these ROIs can contain redundant information and noise, reducing decoding performance. Additionally, the lack of automated ROI labeling methods hinders the practical application of fMRI visual decoding technology, especially for new subjects. This work presents TROI (Trainable Region of Interest), a novel two-stage, data-driven ROI labeling method for cross-subject fMRI decoding tasks, particularly when subject samples are limited. TROI leverages labeled ROIs in the dataset to pretrain an image decoding backbone on a cross-subject dataset, enabling efficient optimization of the input layer for new subjects without retraining the entire model from scratch. In the first stage, we introduce a voxel selection method that combines sparse mask training and low-pass filtering to quickly generate the voxel mask and determine input layer dimensions. In the second stage, we apply a learning rate rewinding strategy to fine-tune the input layer for downstream tasks. Experimental results on the same small sample dataset as the baseline method for brain visual retrieval and reconstruction tasks show that our voxel selection method surpasses the state-of-the-art method MindEye2 with an annotated ROI mask., Comment: ICASSP 2025

Zinc oxide (ZnO) is a promising candidate for hosting point defects as spin qubits for quantum information applications, due to its wide band gap, low spin-orbit coupling, and dilute nuclear spin environment. Previously shallow impurities in ZnO were mostly proposed for spin qubit candidates, but deep-level spin defect studies in ZnO are rather sparse, which may be ideally decoupled from the host materials for stable operation. In this work, we theoretically search for deep-level point defects in ZnO with optimal physical properties suitable for optically-addressable spin qubits in the solid-state. Using first-principles calculations for the search, we have predicted the Mo vacancy defect in ZnO owning promising spin and optical properties, including spin-triplet ground state, optical transition in the visible to near-infrared range with high quantum yield, allowed intersystem crossings with a large optically-detected magnetic resonance contrast, and long spin $T_2$ and $T_2^*$. Notably, we found the Huang-Rhys factor of the defect to be around 5, which is much smaller than those at 10-30 of the most-known defects in ZnO. We also proposed a new protocol for initializing and reading spin qubits, which could be applied in other systems with forbidden longitudinal intersystem crossing. Finally, we compared the spin decoherence driven by the nuclear spin bath and by paramagnetic impurity baths. We found that the paramagnetic impurities are very effective in causing spin decoherence even with very low concentrations, implying that the spin decoherence in ZnO can be likely dominated by them even after isotopic purification.

Emotions play a crucial role in human behavior and decision-making, making emotion recognition a key area of interest in human-computer interaction (HCI). This study addresses the challenges of emotion recognition by integrating facial expression analysis with electroencephalogram (EEG) signals, introducing a novel multimodal framework-Milmer. The proposed framework employs a transformer-based fusion approach to effectively integrate visual and physiological modalities. It consists of an EEG preprocessing module, a facial feature extraction and balancing module, and a cross-modal fusion module. To enhance visual feature extraction, we fine-tune a pre-trained Swin Transformer on emotion-related datasets. Additionally, a cross-attention mechanism is introduced to balance token representation across modalities, ensuring effective feature integration. A key innovation of this work is the adoption of a multiple instance learning (MIL) approach, which extracts meaningful information from multiple facial expression images over time, capturing critical temporal dynamics often overlooked in previous studies. Extensive experiments conducted on the DEAP dataset demonstrate the superiority of the proposed framework, achieving a classification accuracy of 96.72% in the four-class emotion recognition task. Ablation studies further validate the contributions of each module, highlighting the significance of advanced feature extraction and fusion strategies in enhancing emotion recognition performance. Our code are available at https://github.com/liangyubuaa/Milmer.

We study primal-dual algorithms for general empirical risk minimization problems in distributed settings, focusing on two prominent classes of algorithms. The first class is the communication-efficient distributed dual coordinate ascent (CoCoA), derived from the coordinate ascent method for solving the dual problem. The second class is the alternating direction method of multipliers (ADMM), including consensus ADMM, linearized ADMM, and proximal ADMM. We demonstrate that both classes of algorithms can be transformed into a unified update form that involves only primal and dual variables. This discovery reveals key connections between the two classes of algorithms: CoCoA can be interpreted as a special case of proximal ADMM for solving the dual problem, while consensus ADMM is closely related to a proximal ADMM algorithm. This discovery provides the insight that by adjusting the augmented Lagrangian parameter, we can easily enable the ADMM variants to outperform the CoCoA variants. We further explore linearized versions of ADMM and analyze the effects of tuning parameters on these ADMM variants in the distributed setting. Our theoretical findings are supported by extensive simulation studies and real-world data analysis., Comment: 15 pages, 4 figures, 1 table

We present a correlation between the stellar metallicities and the mutual inclinations of multi-planet systems hosting short-period small planets (a/Rs<12, Rp<4Re). We analyzed 89 multi-planet systems discovered by Kepler, K2, and TESS, where the innermost planets have periods shorter than 10 days. We found that the mutual inclinations of the innermost two planets are higher and more diverse around metal-rich stars. The mutual inclinations are calculated as the absolute differences between the best-fit inclinations of the innermost two planets from transit modeling, which represent the lower limits of the true mutual inclinations. The mean and variance of the mutual inclination distribution of the metal-rich systems are 3.1+-0.5 and 3.1+-0.4 degrees, while for the metal-poor systems they are 1.3+-0.2 and 1.0+-0.2 degrees. This finding suggests that inner planetary systems around metal-rich stars are dynamically hotter. We summarized the theories that could plausibly explain this correlation, including the influence of giant planets, higher solid densities in protoplanetary disks around metal-rich stars, or secular chaos coupled with an excess of angular momentum deficits. Planet formation and population synthesis models tracking the mutual inclination evolution would be essential to fully understand this correlation., Comment: 14 pages, 5 figures, 1 table. Accepted by ApJL

Equivariant Graph Neural Networks (EGNNs) have demonstrated significant success in modeling microscale systems, including those in chemistry, biology and materials science. However, EGNNs face substantial computational challenges due to the high cost of constructing edge features via spherical tensor products, making them impractical for large-scale systems. To address this limitation, we introduce E2Former, an equivariant and efficient transformer architecture that incorporates the Wigner $6j$ convolution (Wigner $6j$ Conv). By shifting the computational burden from edges to nodes, the Wigner $6j$ Conv reduces the complexity from $O(|\mathcal{E}|)$ to $ O(| \mathcal{V}|)$ while preserving both the model's expressive power and rotational equivariance. We show that this approach achieves a 7x-30x speedup compared to conventional $\mathrm{SO}(3)$ convolutions. Furthermore, our empirical results demonstrate that the derived E2Former mitigates the computational challenges of existing approaches without compromising the ability to capture detailed geometric information. This development could suggest a promising direction for scalable and efficient molecular modeling.

The capability of GenAI-based chatbots, such as ChatGPT and Gemini, has expanded quickly in recent years, turning them into GenAI Chatbot Ecosystems. Yet, users' understanding of how such ecosystems work remains unknown. In this paper, we investigate users' mental models of how GenAI Chatbot Ecosystems work. This is an important question because users' mental models guide their behaviors, including making decisions that impact their privacy. Through 21 semi-structured interviews, we uncovered users' four mental models towards first-party (e.g., Google Gemini) and third-party (e.g., ChatGPT) GenAI Chatbot Ecosystems. These mental models centered around the role of the chatbot in the entire ecosystem. We further found that participants held a more consistent and simpler mental model towards third-party ecosystems than the first-party ones, resulting in higher trust and fewer concerns towards the third-party ecosystems. We discuss the design and policy implications based on our results.

Chain-of-Thought (CoT) reasoning has become a powerful framework for improving complex problem-solving capabilities in Multimodal Large Language Models (MLLMs). However, the verbose nature of textual reasoning introduces significant inefficiencies. In this work, we propose $\textbf{Heima}$ (as hidden llama), an efficient reasoning framework that leverages reasoning CoTs at hidden latent space. We design the Heima Encoder to condense each intermediate CoT into a compact, higher-level hidden representation using a single thinking token, effectively minimizing verbosity and reducing the overall number of tokens required during the reasoning process. Meanwhile, we design corresponding Heima Decoder with traditional Large Language Models (LLMs) to adaptively interpret the hidden representations into variable-length textual sequence, reconstructing reasoning processes that closely resemble the original CoTs. Experimental results across diverse reasoning MLLM benchmarks demonstrate that Heima model achieves higher generation efficiency while maintaining or even better zero-shot task accuracy. Moreover, the effective reconstruction of multimodal reasoning processes with Heima Decoder validates both the robustness and interpretability of our approach., Comment: Preprint version

Human Action Recognition (HAR) plays a crucial role in applications such as health monitoring, smart home automation, and human-computer interaction. While HAR has been extensively studied, action summarization, which involves identifying and summarizing continuous actions, remains an emerging task. This paper introduces the novel XRF V2 dataset, designed for indoor daily activity Temporal Action Localization (TAL) and action summarization. XRF V2 integrates multimodal data from Wi-Fi signals, IMU sensors (smartphones, smartwatches, headphones, and smart glasses), and synchronized video recordings, offering a diverse collection of indoor activities from 16 volunteers across three distinct environments. To tackle TAL and action summarization, we propose the XRFMamba neural network, which excels at capturing long-term dependencies in untrimmed sensory sequences and outperforms state-of-the-art methods, such as ActionFormer and WiFiTAD. We envision XRF V2 as a valuable resource for advancing research in human action localization, action forecasting, pose estimation, multimodal foundation models pre-training, synthetic data generation, and more., Comment: 27 pages, 11 figures, 8 tables

In this study, we investigate the $CP$-averaged branching fractions and direct $CP$ asymmetries in the quasi-two-body decays $B\to P (f_0(980) \to )(K^+K^-, \pi^+\pi^-)$ within the framework of perturbative QCD approach, where $P$ denotes light pseudoscalar mesons ($\pi$,$K$). The timelike form factor $F_{S}(\omega)$, characterizing final-state interactions between collinear particles in the resonant region, is modeled through the revised relativistic Breit-Wigner formalism for the $S$-wave $f_0(980)$ resonance. Within the Gegenbauer moments of the $S$-wave two-meson distribution amplitudes constrained by current data, we predict the branching fractions of $B\to P (f_0(980) \to )(K^+K^-, \pi^+\pi^-)$ to lie in the range of $10^{-8}$ to $10^{-6}$. Our calculations reveal that our theoretical results for some decay modes exhibit satisfactory agreement with existing measurements from the BaBar and Belle collaborations. Furthermore, we extract the corresponding branching fractions of $B \to P f_0(980)$ decays from the quasi-two-body counterparts $B\to P (f_0(980) \to )\pi^+\pi^-$ rather than $B\to P (f_0(980) \to )K^+K^-$. All predictions are awaiting experimental verification in current and upcoming collider experiments., Comment: 22 pages and 1 figures

We discuss our transient search for directed energy systems in local galaxies, with calculations indicating the ability of modest searches to detect optical Search for Extraterrestrial Intelligence (SETI) sources in the closest galaxies. Our analysis follows Lubin (2016) where a messenger civilization follows a beacon strategy we call "intelligent targeting." We plot the required laser time to achieve an SNR of 10 and find the time for a blind transmission to target all stars in the Milky Way to be achievable for local galactic civilizations. As high cadence and sky coverage is the pathway to enable such a detection, we operate the Local Galactic Transient Survey (LGTS) targeting M31 (the Andromeda Galaxy), the Large Magellanic Cloud (LMC), and the Small Magellanic Cloud (SMC) via Las Cumbres Observatory's (LCO) network of 0.4 m telescopes. We explore the ability of modest searches like the LGTS to detect directed pulses in optical and near-infrared wavelengths from Extraterrestrial Intelligence (ETI) at these distances and conclude a civilization utilizing less powerful laser technology than we can construct in this century is readily detectable with the LGTS's observational capabilities. Data processing of 30,000 LGTS images spanning 5 years is in progress with the TRansient Image Processing Pipeline (TRIPP; Thomas et al. (2025))., Comment: 8 pages, 7 figures, 1 table

Federated Learning (FL) enables clients to collaboratively train machine learning models without sharing local data, preserving privacy in diverse environments. While traditional FL approaches preserve privacy, they often struggle with high computational and communication overhead. To address these issues, model pruning is introduced as a strategy to streamline computations. However, existing pruning methods, when applied solely based on local data, often produce sub-models that inadequately reflect clients' specific tasks due to data insufficiency. To overcome these challenges, this paper introduces SAFL (Structure-Aware Federated Learning), a novel framework that enhances personalized federated learning through client-specific clustering and Similar Client Structure Information (SCSI)-guided model pruning. SAFL employs a two-stage process: initially, it groups clients based on data similarities and uses aggregated pruning criteria to guide the pruning process, facilitating the identification of optimal sub-models. Subsequently, clients train these pruned models and engage in server-based aggregation, ensuring tailored and efficient models for each client. This method significantly reduces computational overhead while improving inference accuracy. Extensive experiments demonstrate that SAFL markedly diminishes model size and improves performance, making it highly effective in federated environments characterized by heterogeneous data.

Large language models (LLMs) such as OpenAI's o1 have demonstrated remarkable abilities in complex reasoning tasks by scaling test-time compute and exhibiting human-like deep thinking. However, we identify a phenomenon we term underthinking, where o1-like LLMs frequently switch between different reasoning thoughts without sufficiently exploring promising paths to reach a correct solution. This behavior leads to inadequate depth of reasoning and decreased performance, particularly on challenging mathematical problems. To systematically analyze this issue, we conduct experiments on three challenging test sets and two representative open-source o1-like models, revealing that frequent thought switching correlates with incorrect responses. We introduce a novel metric to quantify underthinking by measuring token efficiency in incorrect answers. To address underthinking, we propose a decoding strategy with thought switching penalty TIP that discourages premature transitions between thoughts, encouraging deeper exploration of each reasoning path. Experimental results demonstrate that our approach improves accuracy across challenging datasets without requiring model fine-tuning. Our findings contribute to understanding reasoning inefficiencies in o1-like LLMs and offer a practical solution to enhance their problem-solving capabilities., Comment: 1. We have updated the results for DeepSeek-R1, and all of our original conclusions remain valid. 2. Our proposed Tip approach remains effective in Best-of-N scenarios (e.g., self-consistency and Laconic Decoding) when built on DeepSeek-R1

Large Language Models (LLMs) are pretrained on extensive multilingual corpora to acquire both language-specific cultural knowledge and general knowledge. Ideally, while LLMs should provide consistent responses to culture-independent questions across languages, we observe significant performance disparities. To address this, we explore the Cross-Lingual Self-Aligning ability of Language Models (CALM) to align knowledge across languages. Specifically, for a given question, we sample multiple responses across different languages and select the most self-consistent response as the target, leaving the remaining responses as negative examples. We then employ direct preference optimization (DPO) to align the model's knowledge across different languages. Evaluations on the MEDQA and X-CSQA datasets demonstrate CALM's effectiveness in enhancing cross-lingual knowledge question answering, both in zero-shot and retrieval-augmented settings. We also found that increasing the number of languages involved in CALM training leads to higher accuracy and consistency. We offer a qualitative analysis of how cross-lingual consistency can enhance knowledge alignment and explore the method's generalizability., Comment: Accepted by NAACL 2025

Graph Neural Networks (GNNs) have demonstrated remarkable effectiveness in various graph representation learning tasks. However, most existing GNNs focus primarily on capturing local information through explicit graph convolution, often neglecting global message-passing. This limitation hinders the establishment of a collaborative interaction between global and local information, which is crucial for comprehensively understanding graph data. To address these challenges, we propose a novel framework called Comprehensive Graph Representation Learning (ComGRL). ComGRL integrates local information into global information to derive powerful representations. It achieves this by implicitly smoothing local information through flexible graph contrastive learning, ensuring reliable representations for subsequent global exploration. Then ComGRL transfers the locally derived representations to a multi-head self-attention module, enhancing their discriminative ability by uncovering diverse and rich global correlations. To further optimize local information dynamically under the self-supervision of pseudo-labels, ComGRL employs a triple sampling strategy to construct mixed node pairs and applies reliable Mixup augmentation across attributes and structure for local contrastive learning. This approach broadens the receptive field and facilitates coordination between local and global representation learning, enabling them to reinforce each other. Experimental results across six widely used graph datasets demonstrate that ComGRL achieves excellent performance in node classification tasks. The code could be available at https://github.com/JinluWang1002/ComGRL., Comment: 9 pages, 2 figures

Delivering superior search services is crucial for enhancing customer experience and driving revenue growth. Conventionally, search systems model user behaviors by combining user preference and query item relevance statically, often through a fixed logical 'and' relationship. This paper reexamines existing approaches through a unified lens using both causal graphs and Venn diagrams, uncovering two prevalent yet significant issues: entangled preference and relevance effects, and a collapsed modeling space. To surmount these challenges, our research introduces a novel framework, DRP, which enhances search accuracy through two components to reconstruct the behavior modeling space. Specifically, we implement preference editing to proactively remove the relevance effect from preference predictions, yielding untainted user preferences. Additionally, we employ adaptive fusion, which dynamically adjusts fusion criteria to align with the varying patterns of relevance and preference, facilitating more nuanced and tailored behavior predictions within the reconstructed modeling space. Empirical validation on two public datasets and a proprietary search dataset underscores the superiority of our proposed methodology, demonstrating marked improvements in performance over existing approaches.

In autonomous driving, The perception capabilities of the ego-vehicle can be improved with roadside sensors, which can provide a holistic view of the environment. However, existing monocular detection methods designed for vehicle cameras are not suitable for roadside cameras due to viewpoint domain gaps. To bridge this gap and Improve ROAdside Monocular 3D object detection, we propose IROAM, a semantic-geometry decoupled contrastive learning framework, which takes vehicle-side and roadside data as input simultaneously. IROAM has two significant modules. In-Domain Query Interaction module utilizes a transformer to learn content and depth information for each domain and outputs object queries. Cross-Domain Query Enhancement To learn better feature representations from two domains, Cross-Domain Query Enhancement decouples queries into semantic and geometry parts and only the former is used for contrastive learning. Experiments demonstrate the effectiveness of IROAM in improving roadside detector's performance. The results validate that IROAM has the capabilities to learn cross-domain information., Comment: 7 pages, 5 figures, ICRA2025

We present the TRansient Image Processing Pipeline (TRIPP), a transient and variable source detection pipeline that employs both difference imaging and light curve analysis techniques for astronomical data. Additionally, we demonstrate TRIPP's rapid analysis capability by detecting transient candidates in near-real time. TRIPP was tested using image data of the supernova SN2023ixf and from the Local Galactic Transient Survey (LGTS, Thomas et al. (2025)) collected by the Las Cumbres Observatory's (LCO) network of 0.4 m telescopes. To verify the methods employed by TRIPP, we compare our results to published findings on the photometry of SN2023ixf. Additionally, we report the ability of TRIPP to detect transient signals from optical Search for Extra Terrestrial Intelligence (SETI) sources., Comment: 11 pages, 10 figures, 2 tables

Dynamic effective connectivity networks (dECNs) reveal the changing directed brain activity and the dynamic causal influences among brain regions, which facilitate the identification of individual differences and enhance the understanding of human brain. Although the existing causal discovery methods have shown promising results in effective connectivity network analysis, they often overlook the dynamics of causality, in addition to the incorporation of spatio-temporal information in brain activity data. To address these issues, we propose a deep spatio-temporal fusion architecture, which employs a dynamic causal deep encoder to incorporate spatio-temporal information into dynamic causality modeling, and a dynamic causal deep decoder to verify the discovered causality. The effectiveness of the proposed method is first illustrated with simulated data. Then, experimental results from Philadelphia Neurodevelopmental Cohort (PNC) demonstrate the superiority of the proposed method in inferring dECNs, which reveal the dynamic evolution of directed flow between brain regions. The analysis shows the difference of dECNs between young adults and children. Specifically, the directed brain functional networks transit from fluctuating undifferentiated systems to more stable specialized networks as one grows. This observation provides further evidence on the modularization and adaptation of brain networks during development, leading to higher cognitive abilities observed in young adults.

Large Language Models (LLMs) have demonstrated remarkable capabilities in complex reasoning tasks, yet generating reliable reasoning processes remains a significant challenge. We present a unified probabilistic framework that formalizes LLM reasoning through a novel graphical model incorporating latent thinking processes and evaluation signals. Within this framework, we introduce the Bootstrapping Reinforced Thinking Process (BRiTE) algorithm, which works in two steps. First, it generates high-quality rationales by approximating the optimal thinking process through reinforcement learning, using a novel reward shaping mechanism. Second, it enhances the base LLM by maximizing the joint probability of rationale generation with respect to the model's parameters. Theoretically, we demonstrate BRiTE's convergence at a rate of $1/T$ with $T$ representing the number of iterations. Empirical evaluations on math and coding benchmarks demonstrate that our approach consistently improves performance across different base models without requiring human-annotated thinking processes. In addition, BRiTE demonstrates superior performance compared to existing algorithms that bootstrap thinking processes use alternative methods such as rejection sampling, and can even match or exceed the results achieved through supervised fine-tuning with human-annotated data.

Large language models based Multi Agent Systems (MAS) have demonstrated promising performance for enhancing the efficiency and accuracy of code generation tasks. However,most existing methods follow a conventional sequence of planning, coding, and debugging,which contradicts the growth-driven nature of human learning process. Additionally,the frequent information interaction between multiple agents inevitably involves high computational costs. In this paper,we propose Cogito,a neurobiologically inspired multi-agent framework to enhance the problem-solving capabilities in code generation tasks with lower cost. Specifically,Cogito adopts a reverse sequence: it first undergoes debugging, then coding,and finally planning. This approach mimics human learning and development,where knowledge is acquired progressively. Accordingly,a hippocampus-like memory module with different functions is designed to work with the pipeline to provide quick retrieval in similar tasks. Through this growth-based learning model,Cogito accumulates knowledge and cognitive skills at each stage,ultimately forming a Super Role an all capable agent to perform the code generation task. Extensive experiments against representative baselines demonstrate the superior performance and efficiency of Cogito. The code is publicly available at https://anonymous.4open.science/r/Cogito-0083.