Search

DiT diffusion models have achieved great success in text-to-video generation, leveraging their scalability in model capacity and data scale. High content and motion fidelity aligned with text prompts, however, often require large model parameters and a substantial number of function evaluations (NFEs). Realistic and visually appealing details are typically reflected in high resolution outputs, further amplifying computational demands especially for single stage DiT models. To address these challenges, we propose a novel two stage framework, FlashVideo, which strategically allocates model capacity and NFEs across stages to balance generation fidelity and quality. In the first stage, prompt fidelity is prioritized through a low resolution generation process utilizing large parameters and sufficient NFEs to enhance computational efficiency. The second stage establishes flow matching between low and high resolutions, effectively generating fine details with minimal NFEs. Quantitative and visual results demonstrate that FlashVideo achieves state-of-the-art high resolution video generation with superior computational efficiency. Additionally, the two-stage design enables users to preview the initial output and accordingly adjust the prompt before committing to full-resolution generation, thereby significantly reducing computational costs and wait times as well as enhancing commercial viability., Comment: Model and Weight: https://github.com/FoundationVision/FlashVideo

Chest X-ray (CXR) is the most frequently ordered imaging test, supporting diverse clinical tasks from thoracic disease detection to postoperative monitoring. However, task-specific classification models are limited in scope, require costly labeled data, and lack generalizability to out-of-distribution datasets. To address these challenges, we introduce CheXFound, a self-supervised vision foundation model that learns robust CXR representations and generalizes effectively across a wide range of downstream tasks. We pretrain CheXFound on a curated CXR-1M dataset, comprising over one million unique CXRs from publicly available sources. We propose a Global and Local Representations Integration (GLoRI) module for downstream adaptations, by incorporating disease-specific local features with global image features for enhanced performance in multilabel classification. Our experimental results show that CheXFound outperforms state-of-the-art models in classifying 40 disease findings across different prevalence levels on the CXR-LT 24 dataset and exhibits superior label efficiency on downstream tasks with limited training data. Additionally, CheXFound achieved significant improvements on new tasks with out-of-distribution datasets, including opportunistic cardiovascular disease risk estimation and mortality prediction. These results highlight CheXFound's strong generalization capabilities, enabling diverse adaptations with improved label efficiency. The project source code is publicly available at https://github.com/RPIDIAL/CheXFound.

This paper introduces Goku, a state-of-the-art family of joint image-and-video generation models leveraging rectified flow Transformers to achieve industry-leading performance. We detail the foundational elements enabling high-quality visual generation, including the data curation pipeline, model architecture design, flow formulation, and advanced infrastructure for efficient and robust large-scale training. The Goku models demonstrate superior performance in both qualitative and quantitative evaluations, setting new benchmarks across major tasks. Specifically, Goku achieves 0.76 on GenEval and 83.65 on DPG-Bench for text-to-image generation, and 84.85 on VBench for text-to-video tasks. We believe that this work provides valuable insights and practical advancements for the research community in developing joint image-and-video generation models., Comment: Demo: https://saiyan-world.github.io/goku/

Nowadays robot is supposed to demonstrate human-like perception, reasoning and behavior pattern in social or service application. However, most of the existing motion planning methods are incompatible with above requirement. A potential reason is that the existing navigation algorithms usually intend to treat people as another kind of obstacle, and hardly take the social principle or awareness into consideration. In this paper, we attempt to model the proxemics of group and blend it into the scenario perception and navigation of robot. For this purpose, a group clustering method considering both social relevance and spatial confidence is introduced. It can enable robot to identify individuals and divide them into groups. Next, we propose defining the individual proxemics within magnetic dipole model, and further established the group proxemics and scenario map through vector-field superposition. On the basis of the group clustering and proxemics modeling, we present the method to obtain the optimal observation positions (OOPs) of group. Once the OOPs grid and scenario map are established, a heuristic path is employed to generate path that guide robot cruising among the groups for interactive purpose. A series of experiments are conducted to validate the proposed methodology on the practical robot, the results have demonstrated that our methodology has achieved promising performance on group recognition accuracy and path-generation efficiency. This concludes that the group awareness evolved as an important module to make robot socially behave in the practical scenario., Comment: 14 pages,14 figures

A direct measurement of the ground-state-to-ground-state electron-capture decay $Q$ ($Q_{\rm EC}$) value of $^{97}$Tc has been conducted employing the high resolving power phase-imaging ion-cyclotron-resonance technique with the double Penning trap mass spectrometer JYFLTRAP. The resulting $Q_{\rm EC}$ value for $^{97}$Tc is 324.82(21) keV, exhibiting a precision approximately 19 times higher than the value adopted in the newest Atomic Mass Evaluation (AME2020) and differing by 1.2$\sigma$. Furthermore, by combining this refined $Q$ value with nuclear energy-level data for the decay-daughter $^{97}$Mo, a potential ultra-low Q-value transition, possibly of allowed type, $^{97}$Tc (9/2$^{+}$, ground state) $\rightarrow$ $^{97}$Mo$^{*}$ (320(1) keV), was evaluated for future long-term neutrino-mass determination experiments. The ground-state-to-excited-state electron-capture decay $Q$ value ($Q^{*}_{\rm EC}$) of this transition was determined to be 4.8(10) keV, confirming it to be energetically allowed with a confidence level of exceeding 4$\sigma$. The captures of electrons occupying the L and higher shells for this transition are energetically allowed, giving a value of 2.0(10) keV for the closest distance of $Q^{*}_{\rm EC}$ to the allowed binding energy of the L1 shell. To predict partial half-lives and energy-release distributions for this transition, the atomic self-consistent many-electron Dirac--Hartree--Fock--Slater method and the nuclear shell model have been employed. Dominant correction terms such as exchange and overlap corrections, as well as shake-up and shake-off effects, were included in the final results. Moreover, the normalized distribution of released energy in the electron-capture decay of $^{97}$Tc to excited states of $^{97}$Mo, is compared with that of $^{163}$Ho, which is being used for electron-neutrino-mass determination., Comment: 10 pages, 5 figures

As a natural multimodal content, audible video delivers an immersive sensory experience. Consequently, audio-video generation systems have substantial potential. However, existing diffusion-based studies mainly employ relatively independent modules for generating each modality, which lack exploration of shared-weight generative modules. This approach may under-use the intrinsic correlations between audio and visual modalities, potentially resulting in sub-optimal generation quality. To address this, we propose UniForm, a unified diffusion transformer designed to enhance cross-modal consistency. By concatenating auditory and visual information, UniForm learns to generate audio and video simultaneously within a unified latent space, facilitating the creation of high-quality and well-aligned audio-visual pairs. Extensive experiments demonstrate the superior performance of our method in joint audio-video generation, audio-guided video generation, and video-guided audio generation tasks. Our demos are available at https://uniform-t2av.github.io/., Comment: Our demos are available at https://uniform-t2av.github.io/

Nonsymmorphic magnetic Weyl semimetal materials such as ReAlX (Re=rare earth, X=Si/Ge) provide a unique opportunity to explore the correlated phenomena between Weyl fermions and nontrivial magnetic configurations. To be specific, we study a lattice model in which the magnetic configuration is determined by the competition among ferromagnetic (FM) interaction, the Dzyaloshinsky-Moriya interaction, and the Kondo coupling $K_0$ to the Weyl fermion. Both quantum and finite-temperature phase transitions between FM and correlated nesting helical configurations are found. Different from the uncorrelated helimagnet that decouples to the Weyl fermions, this correlated helimagnet induces a magnetic Brillouin zone with a $K_0$-dependent nesting in the band structure of the conducting fermions instead of the magnetic monopole-like Weyl cone. By measuring the current induced by the chiral magnetic effect on the conducting fermion with nesting Weyl nodes, one can distinguish the correlated nesting helical order and the ferromagnetism because the chiral magnetic effect is considerably suppressed in the former case. These properties we study here may explain the experimental observations in ReAlX., Comment: 9 pages, 6 figures

Domain adaptation (DA) for cardiac ultrasound image segmentation is clinically significant and valuable. However, previous domain adaptation methods are prone to be affected by the incomplete pseudo-label and low-quality target to source images. Human-centric domain adaptation has great advantages of human cognitive guidance to help model adapt to target domain and reduce reliance on labels. Doctor gaze trajectories contains a large amount of cross-domain human guidance. To leverage gaze information and human cognition for guiding domain adaptation, we propose gaze-assisted human-centric domain adaptation (GAHCDA), which reliably guides the domain adaptation of cardiac ultrasound images. GAHCDA includes following modules: (1) Gaze Augment Alignment (GAA): GAA enables the model to obtain human cognition general features to recognize segmentation target in different domain of cardiac ultrasound images like humans. (2) Gaze Balance Loss (GBL): GBL fused gaze heatmap with outputs which makes the segmentation result structurally closer to the target domain. The experimental results illustrate that our proposed framework is able to segment cardiac ultrasound images more effectively in the target domain than GAN-based methods and other self-train based methods, showing great potential in clinical application.

This paper introduces IRIS, an immersive Robot Interaction System leveraging Extended Reality (XR), designed for robot data collection and interaction across multiple simulators, benchmarks, and real-world scenarios. While existing XR-based data collection systems provide efficient and intuitive solutions for large-scale data collection, they are often challenging to reproduce and reuse. This limitation arises because current systems are highly tailored to simulator-specific use cases and environments. IRIS is a novel, easily extendable framework that already supports multiple simulators, benchmarks, and even headsets. Furthermore, IRIS is able to include additional information from real-world sensors, such as point clouds captured through depth cameras. A unified scene specification is generated directly from simulators or real-world sensors and transmitted to XR headsets, creating identical scenes in XR. This specification allows IRIS to support any of the objects, assets, and robots provided by the simulators. In addition, IRIS introduces shared spatial anchors and a robust communication protocol that links simulations between multiple XR headsets. This feature enables multiple XR headsets to share a synchronized scene, facilitating collaborative and multi-user data collection. IRIS can be deployed on any device that supports the Unity Framework, encompassing the vast majority of commercially available headsets. In this work, IRIS was deployed and tested on the Meta Quest 3 and the HoloLens 2. IRIS showcased its versatility across a wide range of real-world and simulated scenarios, using current popular robot simulators such as MuJoCo, IsaacSim, CoppeliaSim, and Genesis. In addition, a user study evaluates IRIS on a data collection task for the LIBERO benchmark. The study shows that IRIS significantly outperforms the baseline in both objective and subjective metrics.

This paper presents Perceptual Preference Optimization (PerPO), a perception alignment method aimed at addressing the visual discrimination challenges in generative pre-trained multimodal large language models (MLLMs). To align MLLMs with human visual perception process, PerPO employs discriminative rewarding to gather diverse negative samples, followed by listwise preference optimization to rank them.By utilizing the reward as a quantitative margin for ranking, our method effectively bridges generative preference optimization and discriminative empirical risk minimization. PerPO significantly enhances MLLMs' visual discrimination capabilities while maintaining their generative strengths, mitigates image-unconditional reward hacking, and ensures consistent performance across visual tasks. This work marks a crucial step towards more perceptually aligned and versatile MLLMs. We also hope that PerPO will encourage the community to rethink MLLM alignment strategies.

Fano resonance, arising from the interference between a discrete resonance and a continuum of states, results in sharp and asymmetric line shapes and has significant applications in advanced photonic devices, particularly in sensing, filtering, and nonlinear optics. Nowadays, metasurfaces comprised of engineering microstructures play a crucial role in generation and manipulation of Fano resonance in photonics. However, current metasurfaces dominantly rely on local symmetry breaking of the microstructures to induce Fano resonances, which significant limits their tunability and scalable fabrication for practical applications. To address the challenge, a metal-dielectric hybrid metasurface is demonstrated to achieve nonlocal generation of Fano resonance with no symmetry breaking in the terahertz (THz) band. The Fano resonance, including its existence and peak frequency, is sensitively controlled by the thickness and dielectric constant of the dielectric layer, which is experimentally observed. Our analysis elucidates that the metallic layer with a pair of dumbbell holes leads to the band folding and coupling of guided modes within the dielectric layer. When the thickness or dielectric constant surpasses a critical value, the guided mode resonance falls below the diffraction limit, resulting in a unique nonlocal Fano resonance due to the interaction between the resonance and background transmission facilitated by dumbbell holes. Furthermore, the Fano transmission peak corresponds to an anapole excitation, revealed by multipole calculations. Benefiting from the ability to control the Fano resonance with no symmetry breaking, the proposed hybrid THz metasurface will advance broad applications in the fields of sensors, optical switches, and tunable filters.

The tip of red giant branch (TRGB) stars have attracted intensive attention in recent years because their $I$-band absolute magnitudes, $M_\rm I$, are often used for distance calibration in the Hubble constant measurements because of its almost independence on metallicity ([Fe/H]). However, a discrepancy exists between various studies and the theoretical stellar model predicts dependence of their luminosity on [Fe/H]. Here we present a careful study of the dependence of absolute magnitudes and color indexes on metallicity in optical and near-infrared bands. With the TRGB stars identified in 33 Galactic globular clusters by the reddest color in the $G_{\rm BP}-G_{\rm RP}$ vs. $G_{\rm RP}$ diagram, it is confirmed that $M_\rm I$ is almost constant of $-4.017 \pm 0.036 \pm 0.027$ mag when $[\rm Fe/H]<-1.2$, which would give $H_0=70.86\pm 1.2\pm0.9$ $\rm kms^{-1} Mp c^{-1}$ with this updated luminosity calibration for type Ia supernovae. However, for $[\rm Fe/H]>-1.2$, $M_\rm I$ is found to become fainter with lower metallicity, which would lead to a larger Hubble constant. In the optical $G_{\rm BP}, G_{\rm RP}$ and $V$ bands, the absolute magnitude of TRGB stars tends to increase with metallicity, while in the infrared $J, H$, and $K_{\rm S}$ bands, the variation with metallicity shows an inverse tendency. In addition, the analytical relations of the color indexes with metallicity are presented, which have smaller dispersion than those derived for the corresponding absolute magnitudes.

Maximum entropy reinforcement learning (MaxEnt-RL) has become the standard approach to RL due to its beneficial exploration properties. Traditionally, policies are parameterized using Gaussian distributions, which significantly limits their representational capacity. Diffusion-based policies offer a more expressive alternative, yet integrating them into MaxEnt-RL poses challenges--primarily due to the intractability of computing their marginal entropy. To overcome this, we propose Diffusion-Based Maximum Entropy RL (DIME). DIME leverages recent advances in approximate inference with diffusion models to derive a lower bound on the maximum entropy objective. Additionally, we propose a policy iteration scheme that provably converges to the optimal diffusion policy. Our method enables the use of expressive diffusion-based policies while retaining the principled exploration benefits of MaxEnt-RL, significantly outperforming other diffusion-based methods on challenging high-dimensional control benchmarks. It is also competitive with state-of-the-art non-diffusion based RL methods while requiring fewer algorithmic design choices and smaller update-to-data ratios, reducing computational complexity., Comment: 8 pages main text, 18 pages all included

Machine learning force fields offer the ability to simulate biomolecules with quantum mechanical accuracy while significantly reducing computational costs, attracting growing attention in biophysics. Meanwhile, leveraging the efficiency of molecular mechanics in modeling solvent molecules and long-range interactions, a hybrid machine learning/molecular mechanics (ML/MM) model offers a more realistic approach to describing complex biomolecular systems in solution. However, multiscale models with electrostatic embedding require accounting for the polarization of the ML region induced by the MM environment. To address this, we adapt the state-of-the-art NequIP architecture into a polarizable machine learning force field, NepoIP, enabling the modeling of polarization effects based on the external electrostatic potential. We found that the nanosecond MD simulations based on NepoIP/MM are stable for the periodic solvated dipeptide system and the converged sampling shows excellent agreement with the reference QM/MM level. Moreover, we show that a single NepoIP model can be transferable across different MM force fields, as well as extremely different MM environment of water and proteins, laying the foundation for developing a general machine learning biomolecular force field to be used in ML/MM with electrostatic embedding.

Wearable technology has significantly improved the quality of life for older adults, and the emergence of on-body, movable robots presents new opportunities to further enhance well-being. Yet, the interaction design for these robots remains under-explored, particularly from the perspective of older adults. We present findings from a two-phase co-design process involving 13 older adults to uncover design principles for on-body robots for this population. We identify a rich spectrum of potential applications and characterize a design space to inform how on-body robots should be built for older adults. Our findings highlight the importance of considering factors like co-presence, embodiment, and multi-modal communication. Our work offers design insights to facilitate the integration of on-body robots into daily life and underscores the value of involving older adults in the co-design process to promote usability and acceptance of emerging wearable robotic technologies.

Survival analysis, a vital tool for predicting the time to event, has been used in many domains such as healthcare, criminal justice, and finance. Like classification tasks, survival analysis can exhibit bias against disadvantaged groups, often due to biases inherent in data or algorithms. Several studies in both the IS and CS communities have attempted to address fairness in survival analysis. However, existing methods often overlook the importance of prediction fairness at pre-defined evaluation time points, which is crucial in real-world applications where decision making often hinges on specific time frames. To address this critical research gap, we introduce a new fairness concept: equalized odds (EO) in survival analysis, which emphasizes prediction fairness at pre-defined time points. To achieve the EO fairness in survival analysis, we propose a Conditional Mutual Information Augmentation (CMIA) approach, which features a novel fairness regularization term based on conditional mutual information and an innovative censored data augmentation technique. Our CMIA approach can effectively balance prediction accuracy and fairness, and it is applicable to various survival models. We evaluate the CMIA approach against several state-of-the-art methods within three different application domains, and the results demonstrate that CMIA consistently reduces prediction disparity while maintaining good accuracy and significantly outperforms the other competing methods across multiple datasets and survival models (e.g., linear COX, deep AFT).

Few-shot learning (FSL) has recently been extensively utilized to overcome the scarcity of training data in domain-specific visual recognition. In real-world scenarios, environmental factors such as complex backgrounds, varying lighting conditions, long-distance shooting, and moving targets often cause test images to exhibit numerous incomplete targets or noise disruptions. However, current research on evaluation datasets and methodologies has largely ignored the concept of "environmental robustness", which refers to maintaining consistent performance in complex and diverse physical environments. This neglect has led to a notable decline in the performance of FSL models during practical testing compared to their training performance. To bridge this gap, we introduce a new real-world multi-domain few-shot learning (RD-FSL) benchmark, which includes four domains and six evaluation datasets. The test images in this benchmark feature various challenging elements, such as camouflaged objects, small targets, and blurriness. Our evaluation experiments reveal that existing methods struggle to utilize training images effectively to generate accurate feature representations for challenging test images. To address this problem, we propose a novel conditional representation learning network (CRLNet) that integrates the interactions between training and testing images as conditional information in their respective representation processes. The main goal is to reduce intra-class variance or enhance inter-class variance at the feature representation level. Finally, comparative experiments reveal that CRLNet surpasses the current state-of-the-art methods, achieving performance improvements ranging from 6.83% to 16.98% across diverse settings and backbones. The source code and dataset are available at https://github.com/guoqianyu-alberta/Conditional-Representation-Learning., Comment: 15 pages, 8 figures, Accepted by IEEE Transactions on Image Processing

The transition energies and properties of the Ir$^{17+}$ ion are calculated using the Kramers-restricted configuration-interaction (KRCI) and Fock-space coupled-cluster (FSCC) methods within the Dirac-Coulomb-Gaunt Hamiltonian framework. These calculations show several forbidden optical transitions between the $4f^{13}5s$ ground state and the $4f^{14}$ and $4f^{12}5s^2$ excited states, underscoring their potential as candidates for optical clock applications. Additionally, key properties of the ground and low-lying excited states are reported, including Lande $g_J$ factors, lifetimes, electric dipole polarizabilities, electric quadrupole moments, hyperfine structure constants, relativistic sensitivities, Lorentz-invariance coefficient tensor, and isotope shifts. The excellent agreement between the results from the KRCI and FSCC methods demonstrates the robustness of the calculations and confirms the reliability of the proposed clock transitions., Comment: 10 pages, 4 tables

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

Temporal reasoning in multi-session dialogues presents a significant challenge which has been under-studied in previous temporal reasoning benchmarks. To bridge this gap, we propose a new evaluation task for temporal reasoning in multi-session dialogues and introduce an approach to construct a new benchmark by augmenting dialogues from LoCoMo and creating multi-choice QAs. Furthermore, we present TReMu, a new framework aimed at enhancing the temporal reasoning capabilities of LLM-agents in this context. Specifically, the framework employs \textit{time-aware memorization} through timeline summarization, generating retrievable memory by summarizing events in each dialogue session with their inferred dates. Additionally, we integrate \textit{neuro-symbolic temporal reasoning}, where LLMs generate Python code to perform temporal calculations and select answers. Experimental evaluations on popular LLMs demonstrate that our benchmark is challenging, and the proposed framework significantly improves temporal reasoning performance compared to baseline methods, raising from 29.83 on GPT-4o via standard prompting to 77.67 via our approach and highlighting its effectiveness in addressing temporal reasoning in multi-session dialogues.

The discovery of high-temperature superconductivity in La3Ni2O7 and La4Ni3O10 under high pressure indicates that the Ruddlesden-Popper (RP) phase nickelates Rn+1NinO3n+1 (R = rare earth) is a new material family for high-temperature superconductivity. Exploring the superconductivity of other RP or hybrid RP phase nickelates under high pressure has become an urgent and interesting issue. Here, we report a novel hybrid RP nickelate superconductor of La5Ni3O11. The hybrid RP nickelate La5Ni3O11 is formed by alternative stacking of La3Ni2O7 with n=2 and La2NiO4 with n=1 along the c axis. The transport and magnetic torque measurements indicate a density-wave transition at approximately 170 K near ambient pressure, which is highly similar to both La3Ni2O7 and La4Ni3O10. With increasing pressure, high-pressure transport measurements reveal that the density-wave transition temperature (TDW) continuously increases to approximately 210 K with increasing pressure up to 12 GPa before the appearance of pressure-induced superconductivity, and the density-wave transition abruptly fades out in a first-order manner at approximately 12 GPa. The optimal superconductivity with Tconset = 64 K and Tczero = 54 K is achieved at approximately 21 GPa. On the other hand, high-pressure X-ray diffraction experiments reveal a structural phase transition from an orthorhombic structure to a tetragonal structure at approximately 4.5 GPa. In contrast to La3Ni2O7 and La4Ni3O10, the pressure-induced structural transition has no significant effect on either the density-wave transition or the superconductivity, suggesting a minor role of lattice degree of freedom in La5Ni3O11. The present discovery extends the superconducting member in the RP nickelate family and sheds new light on the superconducting mechanism., Comment: 20 pages, 9 fugures, 1 table

Machine learning (ML) offers considerable promise for the design of new molecules and materials. In real-world applications, the design problem is often domain-specific, and suffers from insufficient data, particularly labeled data, for ML training. In this study, we report a data-efficient, deep-learning framework for molecular discovery that integrates a coarse-grained functional-group representation with a self-attention mechanism to capture intricate chemical interactions. Our approach exploits group-contribution theory to create a graph-based intermediate representation of molecules, serving as a low-dimensional embedding that substantially reduces the data demands typically required for training. By leveraging the self-attention mechanism to learn subtle chemical context, our method consistently outperforms conventional methods in predicting multiple thermophysical properties. In a case study focused on adhesive polymer monomers, we train on a limited dataset comprising just 6,000 unlabeled and 600 labeled monomers. The resulting chemistry prediction model achieves over 92% accuracy in forecasting properties directly from SMILES strings, exceeding the performance of current state-of-the-art techniques. Furthermore, the latent molecular embedding is invertible, allowing the design pipeline to incorporate a decoder that can automatically generate new monomers from the learned chemical subspace. We illustrate this functionality by targeting high and low glass transition temperatures ($T_g$), successfully identifying novel candidates whose $T_g$ extends beyond the range observed in the training data. The ease with which our coarse-grained, attention-based framework navigates both chemical diversity and data scarcity offers a compelling route to accelerate and broaden the search for functional materials.

Node classification is a fundamental task in graph analysis, with broad applications across various fields. Recent breakthroughs in Large Language Models (LLMs) have enabled LLM-based approaches for this task. Although many studies demonstrate the impressive performance of LLM-based methods, the lack of clear design guidelines may hinder their practical application. In this work, we aim to establish such guidelines through a fair and systematic comparison of these algorithms. As a first step, we developed LLMNodeBed, a comprehensive codebase and testbed for node classification using LLMs. It includes ten datasets, eight LLM-based algorithms, and three learning paradigms, and is designed for easy extension with new methods and datasets. Subsequently, we conducted extensive experiments, training and evaluating over 2,200 models, to determine the key settings (e.g., learning paradigms and homophily) and components (e.g., model size) that affect performance. Our findings uncover eight insights, e.g., (1) LLM-based methods can significantly outperform traditional methods in a semi-supervised setting, while the advantage is marginal in a supervised setting; (2) Graph Foundation Models can beat open-source LLMs but still fall short of strong LLMs like GPT-4o in a zero-shot setting. We hope that the release of LLMNodeBed, along with our insights, will facilitate reproducible research and inspire future studies in this field. Codes and datasets are released at \href{https://llmnodebed.github.io/}{https://llmnodebed.github.io/}.

Chronic liver disease represents a significant health challenge worldwide and accurate prognostic evaluations are essential for personalized treatment plans. Recent evidence suggests that integrating multimodal data, such as computed tomography imaging, radiomic features, and clinical information, can provide more comprehensive prognostic information. However, modalities have an inherent heterogeneity, and incorporating additional modalities may exacerbate the challenges of heterogeneous data fusion. Moreover, existing multimodal fusion methods often struggle to adapt to richer medical modalities, making it difficult to capture inter-modal relationships. To overcome these limitations, We present the Triple-Modal Interaction Chronic Liver Network (TMI-CLNet). Specifically, we develop an Intra-Modality Aggregation module and a Triple-Modal Cross-Attention Fusion module, which are designed to eliminate intra-modality redundancy and extract cross-modal information, respectively. Furthermore, we design a Triple-Modal Feature Fusion loss function to align feature representations across modalities. Extensive experiments on the liver prognosis dataset demonstrate that our approach significantly outperforms existing state-of-the-art unimodal models and other multi-modal techniques. Our code is available at https://github.com/Mysterwll/liver.git., Comment: 6 pages, 3 figures, accepted by IEEE ISBI 2025

Employee turnover refers to an individual's termination of employment from the current organization. It is one of the most persistent challenges for firms, especially those ones in Information Technology (IT) industry that confront high turnover rates. Effective prediction of potential employee turnovers benefits multiple stakeholders such as firms and online recruiters. Prior studies have focused on either the turnover prediction within a single firm or the aggregated employee movement among firms. How to predict the individual employees' turnovers among multiple firms has gained little attention in literature, and thus remains a great research challenge. In this study, we propose a novel deep learning approach based on job embeddedness theory to predict the turnovers of individual employees across different firms. Through extensive experimental evaluations using a real-world dataset, our developed method demonstrates superior performance over several state-of-the-art benchmark methods. Additionally, we estimate the cost saving for recruiters by using our turnover prediction solution and interpret the attributions of various driving factors to employee's turnover to showcase its practical business value.

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

Recent advancements in diffusion models have been leveraged to address inverse problems without additional training, and Diffusion Posterior Sampling (DPS) (Chung et al., 2022a) is among the most popular approaches. Previous analyses suggest that DPS accomplishes posterior sampling by approximating the conditional score. While in this paper, we demonstrate that the conditional score approximation employed by DPS is not as effective as previously assumed, but rather aligns more closely with the principle of maximizing a posterior (MAP). This assertion is substantiated through an examination of DPS on 512x512 ImageNet images, revealing that: 1) DPS's conditional score estimation significantly diverges from the score of a well-trained conditional diffusion model and is even inferior to the unconditional score; 2) The mean of DPS's conditional score estimation deviates significantly from zero, rendering it an invalid score estimation; 3) DPS generates high-quality samples with significantly lower diversity. In light of the above findings, we posit that DPS more closely resembles MAP than a conditional score estimator, and accordingly propose the following enhancements to DPS: 1) we explicitly maximize the posterior through multi-step gradient ascent and projection; 2) we utilize a light-weighted conditional score estimator trained with only 100 images and 8 GPU hours. Extensive experimental results indicate that these proposed improvements significantly enhance DPS's performance. The source code for these improvements is provided in https://github.com/tongdaxu/Rethinking-Diffusion-Posterior-Sampling-From-Conditional-Score-Estimator-to-Maximizing-a-Posterior., Comment: ICLR 2025

MRI imputation aims to synthesize the missing modality from one or more available ones, which is highly desirable since it reduces scanning costs and delivers comprehensive MRI information to enhance clinical diagnosis. In this paper, we propose a unified model, CodeBrain, designed to adapt to various brain MRI imputation scenarios. The core design lies in casting various inter-modality transformations as a full-modality code prediction task. To this end, CodeBrain is trained in two stages: Reconstruction and Code Prediction. First, in the Reconstruction stage, we reconstruct each MRI modality, which is mapped into a shared latent space followed by a scalar quantization. Since such quantization is lossy and the code is low dimensional, another MRI modality belonging to the same subject is randomly selected to generate common features to supplement the code and boost the target reconstruction. In the second stage, we train another encoder by a customized grading loss to predict the full-modality codes from randomly masked MRI samples, supervised by the corresponding quantized codes generated from the first stage. In this way, the inter-modality transformation is achieved by mapping the instance-specific codes in a finite scalar space. We evaluated the proposed CodeBrain model on two public brain MRI datasets (i.e., IXI and BraTS 2023). Extensive experiments demonstrate that our CodeBrain model achieves superior imputation performance compared to four existing methods, establishing a new state of the art for unified brain MRI imputation. Codes will be released.

The recent advancements in visual reasoning capabilities of large multimodal models (LMMs) and the semantic enrichment of 3D feature fields have expanded the horizons of robotic capabilities. These developments hold significant potential for bridging the gap between high-level reasoning from LMMs and low-level control policies utilizing 3D feature fields. In this work, we introduce LMM-3DP, a framework that can integrate LMM planners and 3D skill Policies. Our approach consists of three key perspectives: high-level planning, low-level control, and effective integration. For high-level planning, LMM-3DP supports dynamic scene understanding for environment disturbances, a critic agent with self-feedback, history policy memorization, and reattempts after failures. For low-level control, LMM-3DP utilizes a semantic-aware 3D feature field for accurate manipulation. In aligning high-level and low-level control for robot actions, language embeddings representing the high-level policy are jointly attended with the 3D feature field in the 3D transformer for seamless integration. We extensively evaluate our approach across multiple skills and long-horizon tasks in a real-world kitchen environment. Our results show a significant 1.45x success rate increase in low-level control and an approximate 1.5x improvement in high-level planning accuracy compared to LLM-based baselines. Demo videos and an overview of LMM-3DP are available at https://lmm-3dp-release.github.io.

Face animation is a challenging task. Existing model-based methods (utilizing 3DMMs or landmarks) often result in a model-like reconstruction effect, which doesn't effectively preserve identity. Conversely, model-free approaches face challenges in attaining a decoupled and semantically rich feature space, thereby making accurate motion transfer difficult to achieve. We introduce the semantic facial descriptors in learnable disentangled vector space to address the dilemma. The approach involves decoupling the facial space into identity and motion subspaces while endowing each of them with semantics by learning complete orthogonal basis vectors. We obtain basis vector coefficients by employing an encoder on the source and driving faces, leading to effective facial descriptors in the identity and motion subspaces. Ultimately, these descriptors can be recombined as latent codes to animate faces. Our approach successfully addresses the issue of model-based methods' limitations in high-fidelity identity and the challenges faced by model-free methods in accurate motion transfer. Extensive experiments are conducted on three challenging benchmarks (i.e. VoxCeleb, HDTF, CelebV). Comprehensive quantitative and qualitative results demonstrate that our model outperforms SOTA methods with superior identity preservation and motion transfer., Comment: 6 pages,6 figures

Ethical intervention prompting has emerged as a tool to counter demographic biases of text-to-image generative AI models. Existing solutions either require to retrain the model or struggle to generate images that reflect desired distributions on gender and race. We propose an inference-time process called DebiasPI for Debiasing-by-Prompt-Iteration that provides prompt intervention by enabling the user to control the distributions of individuals' demographic attributes in image generation. DebiasPI keeps track of which attributes have been generated either by probing the internal state of the model or by using external attribute classifiers. Its control loop guides the text-to-image model to select not yet sufficiently represented attributes, With DebiasPI, we were able to create images with equal representations of race and gender that visualize challenging concepts of news headlines. We also experimented with the attributes age, body type, profession, and skin tone, and measured how attributes change when our intervention prompt targets the distribution of an unrelated attribute type. We found, for example, if the text-to-image model is asked to balance racial representation, gender representation improves but the skin tone becomes less diverse. Attempts to cover a wide range of skin colors with various intervention prompts showed that the model struggles to generate the palest skin tones. We conducted various ablation studies, in which we removed DebiasPI's attribute control, that reveal the model's propensity to generate young, male characters. It sometimes visualized career success by generating two-panel images with a pre-success dark-skinned person becoming light-skinned with success, or switching gender from pre-success female to post-success male, thus further motivating ethical intervention prompting with DebiasPI., Comment: This work was presented at The European Conference on Computer Vision (ECCV) 2024 Workshop "Fairness and ethics towards transparent AI: facing the chalLEnge through model Debiasing" (FAILED), Milano, Italy, on September 29, 2024, https://failed-workshop-eccv-2024.github.io

Skeleton action recognition models have secured more attention than video-based ones in various applications due to privacy preservation and lower storage requirements. Skeleton data are typically transmitted to cloud servers for action recognition, with results returned to clients via Apps/APIs. However, the vulnerability of skeletal models against adversarial perturbations gradually reveals the unreliability of these systems. Existing black-box attacks all operate in a decision-based manner, resulting in numerous queries that hinder efficiency and feasibility in real-world applications. Moreover, all attacks off the shelf focus on only restricted perturbations, while ignoring model weaknesses when encountered with non-semantic perturbations. In this paper, we propose two query-effIcient Skeletal Adversarial AttaCks, ISAAC-K and ISAAC-N. As a black-box attack, ISAAC-K utilizes Grad-CAM in a surrogate model to extract key joints where minor sparse perturbations are then added to fool the classifier. To guarantee natural adversarial motions, we introduce constraints of both bone length and temporal consistency. ISAAC-K finds stronger adversarial examples on $\ell_\infty$ norm, which can encompass those on other norms. Exhaustive experiments substantiate that ISAAC-K can uplift the attack efficiency of the perturbations under 10 skeletal models. Additionally, as a byproduct, ISAAC-N fools the classifier by replacing skeletons unrelated to the action. We surprisingly find that skeletal models are vulnerable to large perturbations where the part-wise non-semantic joints are just replaced, leading to a query-free no-box attack without any prior knowledge. Based on that, four adaptive defenses are eventually proposed to improve the robustness of skeleton recognition models., Comment: 13 pages, 13 figures

Pulmonary nodules are an early sign of lung cancer, and detecting them early is vital for improving patient survival rates. Most current methods use only single Computed Tomography (CT) images to assess nodule malignancy. However, doctors typically make a comprehensive assessment in clinical practice by integrating follow-up CT scans with clinical data. To enhance this process, our study introduces a Cross-Modal Spatiotemporal Fusion Network, named CSF-Net, designed to predict the malignancy of pulmonary nodules using follow-up CT scans. This approach simulates the decision-making process of clinicians who combine follow-up imaging with clinical information. CSF-Net comprises three key components: spatial feature extraction module, temporal residual fusion module, and cross-modal attention fusion module. Together, these modules enable precise predictions of nodule malignancy. Additionally, we utilized the publicly available NLST dataset to screen and annotate the specific locations of pulmonary nodules and created a new dataset named NLST-cmst. Our experimental results on the NLST-cmst dataset demonstrate significant performance improvements, with an accuracy of 0.8974, a precision of 0.8235, an F1 score of 0.8750, an AUC of 0.9389, and a recall of 0.9333. These findings indicate that our multimodal spatiotemporal fusion approach, which combines follow-up data with clinical information, surpasses existing methods, underscoring its effectiveness in predicting nodule malignancy., Comment: This paper has been accepted by the 2025 IEEE International Symposium on Biomedical Imaging

Bug fixing holds significant importance in software development and maintenance. Recent research has made substantial strides in exploring the potential of large language models (LLMs) for automatically resolving software bugs. However, a noticeable gap in existing approaches lies in the oversight of collaborative facets intrinsic to bug resolution, treating the process as a single-stage endeavor. Moreover, most approaches solely take the buggy code snippet as input for LLMs during the patch generation stage. To mitigate the aforementioned limitations, we introduce a novel stage-wise framework named PATCH. Specifically, we first augment the buggy code snippet with corresponding dependence context and intent information to better guide LLMs in generating the correct candidate patches. Additionally, by taking inspiration from bug management practices, we decompose the bug-fixing task into four distinct stages: bug reporting, bug diagnosis, patch generation, and patch verification. These stages are performed interactively by LLMs, aiming to simulate the collaborative behavior of programmers during the resolution of software bugs. By harnessing these collective contributions, PATCH effectively enhances the bug-fixing capability of LLMs. We implement PATCH by employing the powerful dialogue-based LLM ChatGPT. Our evaluation on the widely used bug-fixing benchmark BFP demonstrates that PATCH has achieved better performance than state-of-the-art LLMs., Comment: Preprint, to appear in the ACM Transactions on Software Engineering and Methodology (TOSEM)

Significant differences in protein structures hinder the generalization of existing drug-target interaction (DTI) models, which often rely heavily on pre-learned binding principles or detailed annotations. In contrast, BioBridge designs an Inductive-Associative pipeline inspired by the workflow of scientists who base their accumulated expertise on drawing insights into novel drug-target pairs from weakly related references. BioBridge predicts novel drug-target interactions using limited sequence data, incorporating multi-level encoders with adversarial training to accumulate transferable binding principles. On these principles basis, BioBridge employs a dynamic prototype meta-learning framework to associate insights from weakly related annotations, enabling robust predictions for previously unseen drug-target pairs. Extensive experiments demonstrate that BioBridge surpasses existing models, especially for unseen proteins. Notably, when only homologous protein binding data is available, BioBridge proves effective for virtual screening of the epidermal growth factor receptor and adenosine receptor, underscoring its potential in drug discovery.

Customizable role-playing in large language models (LLMs), also known as character generalization, is gaining increasing attention for its versatility and cost-efficiency in developing and deploying role-playing dialogue agents. This study explores a large-scale data synthesis approach to equip LLMs with character generalization capabilities. We begin by synthesizing large-scale character profiles using personas from Persona Hub and then explore two strategies: response rewriting and response generation, to create character-aligned instructional responses. To validate the effectiveness of our synthetic instruction tuning data for character generalization, we perform supervised fine-tuning (SFT) using the LLaMA-3 8B model. Our best-performing model strengthens the original LLaMA-3 8B Instruct model and achieves performance comparable to GPT-4o models on role-playing dialogue. We release our synthetic characters and instruction-tuning dialogues to support public research.

Code review remains a critical yet resource-intensive process in software development, particularly challenging in large-scale industrial environments. While Large Language Models (LLMs) show promise for automating code review, existing solutions face significant limitations in precision and practicality. This paper presents BitsAI-CR, an innovative framework that enhances code review through a two-stage approach combining RuleChecker for initial issue detection and ReviewFilter for precision verification. The system is built upon a comprehensive taxonomy of review rules and implements a data flywheel mechanism that enables continuous performance improvement through structured feedback and evaluation metrics. Our approach introduces an Outdated Rate metric that can reflect developers' actual adoption of review comments, enabling automated evaluation and systematic optimization at scale. Empirical evaluation demonstrates BitsAI-CR's effectiveness, achieving 75.0% precision in review comment generation. For the Go language which has predominant usage at ByteDance, we maintain an Outdated Rate of 26.7%. The system has been successfully deployed at ByteDance, serving over 12,000 Weekly Active Users (WAU). Our work provides valuable insights into the practical application of automated code review and offers a blueprint for organizations seeking to implement automated code reviews at scale.

Creating visually appealing advertising images is often a labor-intensive and time-consuming process. Is it possible to automatically generate such images using only basic product information--specifically, a product foreground image, taglines, and a target size? Existing methods mainly focus on parts of the problem and fail to provide a comprehensive solution. To address this gap, we propose a novel multistage framework called Product-Centric Advertising Image Design (PAID). It consists of four sequential stages to highlight product foregrounds and taglines while achieving overall image aesthetics: prompt generation, layout generation, background image generation, and graphics rendering. Different expert models are designed and trained for the first three stages: First, we use a visual language model (VLM) to generate background prompts that match the products. Next, a VLM-based layout generation model arranges the placement of product foregrounds, graphic elements (taglines and decorative underlays), and various nongraphic elements (objects from the background prompt). Following this, we train an SDXL-based image generation model that can simultaneously accept prompts, layouts, and foreground controls. To support the PAID framework, we create corresponding datasets with over 50,000 labeled images. Extensive experimental results and online A/B tests demonstrate that PAID can produce more visually appealing advertising images.

Examining the alignment of large language models (LLMs) has become increasingly important, e.g., when LLMs fail to operate as intended. This study examines the alignment of LLMs with human values for the domain of politics. Prior research has shown that LLM-generated outputs can include political leanings and mimic the stances of political parties on various issues. However, the extent and conditions under which LLMs deviate from empirical positions are insufficiently examined. To address this gap, we analyze the factors that contribute to LLMs' deviations from empirical positions on political issues, aiming to quantify these deviations and identify the conditions that cause them. Drawing on findings from cognitive science about representativeness heuristics, i.e., situations where humans lean on representative attributes of a target group in a way that leads to exaggerated beliefs, we scrutinize LLM responses through this heuristics' lens. We conduct experiments to determine how LLMs inflate predictions about political parties, which results in stereotyping. We find that while LLMs can mimic certain political parties' positions, they often exaggerate these positions more than human survey respondents do. Also, LLMs tend to overemphasize representativeness more than humans. This study highlights the susceptibility of LLMs to representativeness heuristics, suggesting a potential vulnerability of LLMs that facilitates political stereotyping. We also test prompt-based mitigation strategies, finding that strategies that can mitigate representative heuristics in humans are also effective in reducing the influence of representativeness on LLM-generated responses., Comment: ICLR 2025

Recently, emotional speech generation and speaker cloning have garnered significant interest in text-to-speech (TTS). With the open-sourcing of codec language TTS models trained on massive datasets with large-scale parameters, adapting these general pre-trained TTS models to generate speech with specific emotional expressions and target speaker characteristics has become a topic of great attention. Common approaches, such as full and adapter-based fine-tuning, often overlook the specific contributions of model parameters to emotion and speaker control. Treating all parameters uniformly during fine-tuning, especially when the target data has limited content diversity compared to the pre-training corpus, results in slow training speed and an increased risk of catastrophic forgetting. To address these challenges, we propose a characteristic-specific partial fine-tuning strategy, short as CSP-FT. First, we use a weighted-sum approach to analyze the contributions of different Transformer layers in a pre-trained codec language TTS model for emotion and speaker control in the generated speech. We then selectively fine-tune the layers with the highest and lowest characteristic-specific contributions to generate speech with target emotional expression and speaker identity. Experimental results demonstrate that our method achieves performance comparable to, or even surpassing, full fine-tuning in generating speech with specific emotional expressions and speaker identities. Additionally, CSP-FT delivers approximately 2x faster training speeds, fine-tunes only around 8% of parameters, and significantly reduces catastrophic forgetting. Furthermore, we show that codec language TTS models perform competitively with self-supervised models in speaker identification and emotion classification tasks, offering valuable insights for developing universal speech processing models., Comment: 13 pages

An emerging approach for large-scale hydrogen production using renewable energy is to integrate multiple alkaline water electrolysis (AWE) stacks into a single balance of plant (BoP) system, sharing components such as gas-lye separation and lye circulation. This configuration, termed the $N$-in-1 AWE system, packs $N$ stacks into a modular system, reducing land requirements, the complexity of plant topology, and overall capital costs. However, the coupling of these stacks through the shared BoP introduces challenges in dynamic operation under varying energy inputs, making their performance unclear compared to traditional 1-in-1 systems. To address this, we develop a state-space model of the $N$-in-1 AWE system, capturing the dynamic behaviors of lye circulation, temperature, and HTO impurity, and their impact on energy conversion efficiency. We then propose a nonlinear model predictive controller (NMPC) to coordinately optimize inter-stack electrolytic current distribution, lye flow, and cooling, enabling the system to dynamically track varying load commands while maximizing efficiency, stabilizing temperature, and limiting HTO impurity accumulation. Simulation studies on a 4,000 Nm$^3$/h-rated 4-in-1 system verify the proposed controller under dynamic operation. Comparison with 4 independent 1-in-1 systems reveals that, with proper control, the $N$-in-1 configuration offers comparable flexibility in accommodating real-world wind power inputs. The average differences in the root-mean-square errors (RMSEs) for load-tracking and stack temperature stabilization, and specific energy consumption are below 0.014 MW, 2.356 K, and 0.003 kWh/Nm$^3$.

Granular materials undergo compaction under periodic temperature fluctuations, leading to various engineering and geological phenomena from landslides to silo compaction. Although thermal effects on granular materials have been extensively studied in soil mechanics and geology, the underlying physical mechanisms remain unclear. This study investigates the compaction dynamics of granular materials subjected to thermal cycling using monodisperse glass beads and polydisperse sand packings. We demonstrate that differential thermal expansion between the container and the grains drives compaction through shear in our experimental systems. We quantify compaction dynamics using three established fitting models: Kohlrausch-Williams-Watts (KWW), double-exponential, and logarithmic functions. Our results reveal that granular materials exhibit slow relaxation processes in response to weak perturbations, displaying aging dynamics similar to those observed in glassy systems. These findings provide insights into fundamental mechanisms of granular compaction with broad implications for geological and engineering applications., Comment: 19 pages, 6 figures

Penetration testing is a vital practice for identifying and mitigating vulnerabilities in cybersecurity systems, but its manual execution is labor-intensive and time-consuming. Existing large language model (LLM)-assisted or automated penetration testing approaches often suffer from inefficiencies, such as a lack of contextual understanding and excessive, unstructured data generation. This paper presents VulnBot, an automated penetration testing framework that leverages LLMs to simulate the collaborative workflow of human penetration testing teams through a multi-agent system. To address the inefficiencies and reliance on manual intervention in traditional penetration testing methods, VulnBot decomposes complex tasks into three specialized phases: reconnaissance, scanning, and exploitation. These phases are guided by a penetration task graph (PTG) to ensure logical task execution. Key design features include role specialization, penetration path planning, inter-agent communication, and generative penetration behavior. Experimental results demonstrate that VulnBot outperforms baseline models such as GPT-4 and Llama3 in automated penetration testing tasks, particularly showcasing its potential in fully autonomous testing on real-world machines.

As large language models (LLMs) have been deployed in various real-world settings, concerns about the harm they may propagate have grown. Various jailbreaking techniques have been developed to expose the vulnerabilities of these models and improve their safety. This work reveals that many state-of-the-art LLMs are vulnerable to malicious requests hidden behind scientific language. Specifically, our experiments with GPT4o, GPT4o-mini, GPT-4, LLama3-405B-Instruct, Llama3-70B-Instruct, Cohere, Gemini models demonstrate that, the models' biases and toxicity substantially increase when prompted with requests that deliberately misinterpret social science and psychological studies as evidence supporting the benefits of stereotypical biases. Alarmingly, these models can also be manipulated to generate fabricated scientific arguments claiming that biases are beneficial, which can be used by ill-intended actors to systematically jailbreak these strong LLMs. Our analysis studies various factors that contribute to the models' vulnerabilities to malicious requests in academic language. Mentioning author names and venues enhances the persuasiveness of models, and the bias scores increase as dialogues progress. Our findings call for a more careful investigation on the use of scientific data for training LLMs., Comment: 16 pages

The Process Reward Model (PRM) plays a crucial role in mathematical reasoning tasks, requiring high-quality supervised process data. However, we observe that reasoning steps generated by Large Language Models (LLMs) often fail to exhibit strictly incremental information, leading to redundancy that can hinder effective reasoning. To address this issue, we propose CFPRM, a simple yet effective coarse-to-fine strategy. Instead of focusing on the detection of redundant steps, our approach first establishes a coarse-grained window to merge adjacent reasoning steps into unified, holistic steps. The window size is then progressively reduced to extract fine-grained reasoning steps, enabling data collection at multiple granularities for training. By leveraging this hierarchical refinement process, CFPRM mitigates redundancy while preserving essential fine-grained knowledge. Extensive experiments on two reasoning datasets across three loss criteria validate the CFPRM's effectiveness and versatility.

We introduce our first-generation reasoning models, DeepSeek-R1-Zero and DeepSeek-R1. DeepSeek-R1-Zero, a model trained via large-scale reinforcement learning (RL) without supervised fine-tuning (SFT) as a preliminary step, demonstrates remarkable reasoning capabilities. Through RL, DeepSeek-R1-Zero naturally emerges with numerous powerful and intriguing reasoning behaviors. However, it encounters challenges such as poor readability, and language mixing. To address these issues and further enhance reasoning performance, we introduce DeepSeek-R1, which incorporates multi-stage training and cold-start data before RL. DeepSeek-R1 achieves performance comparable to OpenAI-o1-1217 on reasoning tasks. To support the research community, we open-source DeepSeek-R1-Zero, DeepSeek-R1, and six dense models (1.5B, 7B, 8B, 14B, 32B, 70B) distilled from DeepSeek-R1 based on Qwen and Llama.

The recent discovery of superconductivity in La$_3$Ni$_2$O$_7$ and La$_4$Ni$_3$O$_{10}$ under high pressure stimulates intensive research interests. These nickelates crystallize in an orthogonal/monoclinic structure with tilted NiO$_6$ octahedra at ambient pressure and enter a density-wave-like phase at low temperatures. The application of pressure suppresses the octahedral tilting and triggers a transition to tetragonal structure (I4/mmm), which is believed to be a key prerequisite for the emergence of superconducting state. Here, by developing a high oxidative environment growth technology, we report the first tetragonal nickelates La$_4$Ni$_3$O$_{10}$ microcrystals without octahedral tilting at ambient pressure. In tetragonal La$_4$Ni$_3$O$_{10}$, transport measurements find that both density-wave and superconducting transitions are absent up to 160 GPa, indicating a robust tetragonal metallic ground state. Density functional theory calculations reveal that the band structure of ambient-pressure tetragonal La$_4$Ni$_3$O$_{10}$ involves more $d_{z2}$ orbital contribution to the Fermi surface, compared to the monoclinic phase or the high-pressure superconducting tetragonal phase. The concurrent absence of density-wave state and high-pressure superconductivity in our ambient-pressure tetragonal crystals of La$_4$Ni$_3$O$_{10}$ suggests an underlying correlation between these two orders. It suggests that the tetragonal structure is not necessary, while the density-wave state is crucial for the superconductivity in nickelates. Our findings impose important constraints on the mechanism of pressure-induced superconductivity in nickelates and sheds new light on exploring ambient pressure high-temperature Ni-based superconductors., Comment: submitted

We introduce MAGI, a hybrid video generation framework that combines masked modeling for intra-frame generation with causal modeling for next-frame generation. Our key innovation, Complete Teacher Forcing (CTF), conditions masked frames on complete observation frames rather than masked ones (namely Masked Teacher Forcing, MTF), enabling a smooth transition from token-level (patch-level) to frame-level autoregressive generation. CTF significantly outperforms MTF, achieving a +23% improvement in FVD scores on first-frame conditioned video prediction. To address issues like exposure bias, we employ targeted training strategies, setting a new benchmark in autoregressive video generation. Experiments show that MAGI can generate long, coherent video sequences exceeding 100 frames, even when trained on as few as 16 frames, highlighting its potential for scalable, high-quality video generation., Comment: 12 pages, 9 figures

By employing the replica trick we study the impact of the replica parameter $n$ on the modular entropy and the capacity of entanglement in the End of the World (EoW) model and the island model, respectively. For the EoW model, we present $n$-dependent evolution curves of the modular entropy and the capacity of entanglement under both microcanonical and canonical ensembles. In particular, in the canonical ensemble, all quantities decrease as $n$ increases at late times. For the island model, we develop the replica geometry for finite $n$ and re-evaluate the modular entropy and the capacity of entanglement in a two-sided eternal Jackiw-Teitelboim black hole coupled with a thermal bath. In the case of a single island configuration, the modular entropy and capacity of entanglement closely resemble the thermal entropy and the heat capacity, respectively, yielding results analogous to those obtained in the canonical ensemble for the EoW model. The analysis of the results from these two models strongly indicates that in geometries with a greater number of $n$ copies, more connected geometries effectively purify thermal Hawking radiation. In addition, we compare these findings with statistical mechanics and provide an interpretation for the replica parameter $n$. Finally, we generalize the island formula to accommodate the finite $n$ case under this interpretation., Comment: 46 pages, 11 figures, 2 tables

Alzheimer's disease (AD) is a common neurodegenerative disease among the elderly. Early prediction and timely intervention of its prodromal stage, mild cognitive impairment (MCI), can decrease the risk of advancing to AD. Combining information from various modalities can significantly improve predictive accuracy. However, challenges such as missing data and heterogeneity across modalities complicate multimodal learning methods as adding more modalities can worsen these issues. Current multimodal fusion techniques often fail to adapt to the complexity of medical data, hindering the ability to identify relationships between modalities. To address these challenges, we propose an innovative multimodal approach for predicting MCI conversion, focusing specifically on the issues of missing positron emission tomography (PET) data and integrating diverse medical information. The proposed incomplete triple-modal MCI conversion prediction network is tailored for this purpose. Through the missing modal generation module, we synthesize the missing PET data from the magnetic resonance imaging and extract features using specifically designed encoders. We also develop a channel aggregation module and a triple-modal co-attention fusion module to reduce feature redundancy and achieve effective multimodal data fusion. Furthermore, we design a loss function to handle missing modality issues and align cross-modal features. These components collectively harness multimodal data to boost network performance. Experimental results on the ADNI1 and ADNI2 datasets show that our method significantly surpasses existing unimodal and other multimodal models. Our code is available at https://github.com/justinhxy/ITFC., Comment: 5 pages, 1 figure, accepted by IEEE ISBI 2025