Search

Creating Computer-Aided Design (CAD) models requires significant expertise and effort. Text-to-CAD, which converts textual descriptions into CAD parametric sequences, is crucial in streamlining this process. Recent studies have utilized ground-truth parametric sequences, known as sequential signals, as supervision to achieve this goal. However, CAD models are inherently multimodal, comprising parametric sequences and corresponding rendered visual objects. Besides,the rendering process from parametric sequences to visual objects is many-to-one. Therefore, both sequential and visual signals are critical for effective training. In this work, we introduce CADFusion, a framework that uses Large Language Models (LLMs) as the backbone and alternates between two training stages: the sequential learning (SL) stage and the visual feedback (VF) stage. In the SL stage, we train LLMs using ground-truth parametric sequences, enabling the generation of logically coherent parametric sequences. In the VF stage, we reward parametric sequences that render into visually preferred objects and penalize those that do not, allowing LLMs to learn how rendered visual objects are perceived and evaluated. These two stages alternate throughout the training, ensuring balanced learning and preserving benefits of both signals. Experiments demonstrate that CADFusion significantly improves performance, both qualitatively and quantitatively.

Scaling laws are powerful tools to predict the performance of large language models. However, current scaling laws fall short of accounting for inference costs. In this work, we first show that model architecture affects inference latency, where models of the same size can have up to 3.5x difference in latency. To tackle this challenge, we modify the Chinchilla scaling laws to co-optimize the model parameter count, the number of training tokens, and the model architecture. Due to the reason that models of similar training loss exhibit gaps in downstream evaluation, we also propose a novel method to train inference-efficient models based on the revised scaling laws. We perform extensive empirical studies to fit and evaluate our inference-aware scaling laws. We vary model parameters from 80M to 1B, training tokens from 1.6B to 30B, and model shapes, training a total of 63 models. Guided by our inference-efficient scaling law and model selection method, we release the Morph-1B model, which improves inference latency by 1.8x while maintaining accuracy on downstream tasks compared to open-source models, pushing the Pareto frontier of accuracy-latency tradeoff., Comment: 17 pages, 16 figures

Applying Multimodal Large Language Models (MLLMs) to video understanding presents significant challenges due to the need to model temporal relations across frames. Existing approaches adopt either implicit temporal modeling, relying solely on the LLM decoder, or explicit temporal modeling, employing auxiliary temporal encoders. To investigate this debate between the two paradigms, we propose the Stackable Temporal Encoder (STE). STE enables flexible explicit temporal modeling with adjustable temporal receptive fields and token compression ratios. Using STE, we systematically compare implicit and explicit temporal modeling across dimensions such as overall performance, token compression effectiveness, and temporal-specific understanding. We also explore STE's design considerations and broader impacts as a plug-in module and in image modalities. Our findings emphasize the critical role of explicit temporal modeling, providing actionable insights to advance video MLLMs.

Human activity recognition (HAR) ideally relies on data from wearable or environment-instrumented sensors sampled at regular intervals, enabling standard neural network models optimized for consistent time-series data as input. However, real-world sensor data often exhibits irregular sampling due to, for example, hardware constraints, power-saving measures, or communication delays, posing challenges for deployed static HAR models. This study assesses the impact of sampling irregularities on HAR by simulating irregular data through two methods: introducing slight inconsistencies in sampling intervals (timestamp variations) to mimic sensor jitter, and randomly removing data points (random dropout) to simulate missing values due to packet loss or sensor failure. We evaluate both discrete-time neural networks and continuous-time neural networks, which are designed to handle continuous-time data, on three public datasets. We demonstrate that timestamp variations do not significantly affect the performance of discrete-time neural networks, and the continuous-time neural network is also ineffective in addressing the challenges posed by irregular sampling, possibly due to limitations in modeling complex temporal patterns with missing data. Our findings underscore the necessity for new models or approaches that can robustly handle sampling irregularity in time-series data, like the reading in human activity recognition, paving the way for future research in this domain., Comment: Accepted by PerFail (4th International Workshop on Negative Results in Pervasive Computing)

Cross-Domain Sequential Recommendation (CDSR) has recently gained attention for countering data sparsity by transferring knowledge across domains. A common approach merges domain-specific sequences into cross-domain sequences, serving as bridges to connect domains. One key challenge is to correctly extract the shared knowledge among these sequences and appropriately transfer it. Most existing works directly transfer unfiltered cross-domain knowledge rather than extracting domain-invariant components and adaptively integrating them into domain-specific modelings. Another challenge lies in aligning the domain-specific and cross-domain sequences. Existing methods align these sequences based on timestamps, but this approach can cause prediction mismatches when the current tokens and their targets belong to different domains. In such cases, the domain-specific knowledge carried by the current tokens may degrade performance. To address these challenges, we propose the A-B-Cross-to-Invariant Learning Recommender (ABXI). Specifically, leveraging LoRA's effectiveness for efficient adaptation, ABXI incorporates two types of LoRAs to facilitate knowledge adaptation. First, all sequences are processed through a shared encoder that employs a domain LoRA for each sequence, thereby preserving unique domain characteristics. Next, we introduce an invariant projector that extracts domain-invariant interests from cross-domain representations, utilizing an invariant LoRA to adapt these interests into modeling each specific domain. Besides, to avoid prediction mismatches, all domain-specific sequences are aligned to match the domains of the cross-domain ground truths. Experimental results on three datasets demonstrate that our approach outperforms other CDSR counterparts by a large margin. The codes are available in \url{https://github.com/DiMarzioBian/ABXI}., Comment: Accepted by WebConf '25 (WWW '25)

Benchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achieve over 90\% accuracy on popular benchmarks like MMLU, limiting informed measurement of state-of-the-art LLM capabilities. In response, we introduce Humanity's Last Exam (HLE), a multi-modal benchmark at the frontier of human knowledge, designed to be the final closed-ended academic benchmark of its kind with broad subject coverage. HLE consists of 3,000 questions across dozens of subjects, including mathematics, humanities, and the natural sciences. HLE is developed globally by subject-matter experts and consists of multiple-choice and short-answer questions suitable for automated grading. Each question has a known solution that is unambiguous and easily verifiable, but cannot be quickly answered via internet retrieval. State-of-the-art LLMs demonstrate low accuracy and calibration on HLE, highlighting a significant gap between current LLM capabilities and the expert human frontier on closed-ended academic questions. To inform research and policymaking upon a clear understanding of model capabilities, we publicly release HLE at https://lastexam.ai., Comment: 25 pages, 6 figures

Low-rank tensor completion aims to recover a tensor from partially observed entries, and it is widely applicable in fields such as quantum computing and image processing. Due to the significant advantages of the tensor train (TT) format in handling structured high-order tensors, this paper investigates the low-rank tensor completion problem based on the TT-format. We proposed a preconditioned Riemannian gradient descent algorithm (PRGD) to solve low TT-rank tensor completion and establish its linear convergence. Experimental results on both simulated and real datasets demonstrate the effectiveness of the PRGD algorithm. On the simulated dataset, the PRGD algorithm reduced the computation time by two orders of magnitude compared to existing classical algorithms. In practical applications such as hyperspectral image completion and quantum state tomography, the PRGD algorithm significantly reduced the number of iterations, thereby substantially reducing the computational time.

Graph computational tasks are inherently challenging and often demand the development of advanced algorithms for effective solutions. With the emergence of large language models (LLMs), researchers have begun investigating their potential to address these tasks. However, existing approaches are constrained by LLMs' limited capability to comprehend complex graph structures and their high inference costs, rendering them impractical for handling large-scale graphs. Inspired by human approaches to graph problems, we introduce a novel framework, PIE (Pseudocode-Injection-Enhanced LLM Reasoning for Graph Computational Tasks), which consists of three key steps: problem understanding, prompt design, and code generation. In this framework, LLMs are tasked with understanding the problem and extracting relevant information to generate correct code. The responsibility for analyzing the graph structure and executing the code is delegated to the interpreter. We inject task-related pseudocodes into the prompts to further assist the LLMs in generating efficient code. We also employ cost-effective trial-and-error techniques to ensure that the LLM-generated code executes correctly. Unlike other methods that require invoking LLMs for each individual test case, PIE only calls the LLM during the code generation phase, allowing the generated code to be reused and significantly reducing inference costs. Extensive experiments demonstrate that PIE outperforms existing baselines in terms of both accuracy and computational efficiency., Comment: 24 pages

The KOTO II experiment is proposed to measure the branching ratio of the decay $K_L\to\pi^0\nu\bar{\nu}$ at J-PARC. With a beamline to extract long-lived neutral kaons at 5 degrees from a production target, the single event sensitivity of the decay is $8.5\times 10^{-13}$, which is much smaller than the Standard Model prediction $3\times 10^{-11}$. This allows searches for new physics beyond the Standard Model and the first discovery of the decay with a significance exceeding $5\sigma$. As the only experiment proposed in the world dedicated to rare kaon decays, KOTO II will be indispensable in the quest for a complete understanding of flavor dynamics in the quark sector. Moreover, by combining efforts from the kaon community worldwide, we plan to develop the KOTO II detector further and expand the physics reach of the experiment to include measurements of the branching ratio of the $K_L\to\pi^0\ell^+\ell^-$ decays, studies of other $K_L$ decays, and searches for dark photons, axions, and axion-like particles. KOTO II will therefore obtain a comprehensive understanding of $K_L$ decays, providing further constraints on new physics scenarios with existing $K^+$ results., Comment: Submitted to the J-PARC PAC. arXiv admin note: substantial text overlap with arXiv:2110.04462

Deep learning based person re-identification (re-id) models have been widely employed in surveillance systems. Recent studies have demonstrated that black-box single-modality and cross-modality re-id models are vulnerable to adversarial examples (AEs), leaving the robustness of multi-modality re-id models unexplored. Due to the lack of knowledge about the specific type of model deployed in the target black-box surveillance system, we aim to generate modality unified AEs for omni-modality (single-, cross- and multi-modality) re-id models. Specifically, we propose a novel Modality Unified Attack method to train modality-specific adversarial generators to generate AEs that effectively attack different omni-modality models. A multi-modality model is adopted as the surrogate model, wherein the features of each modality are perturbed by metric disruption loss before fusion. To collapse the common features of omni-modality models, Cross Modality Simulated Disruption approach is introduced to mimic the cross-modality feature embeddings by intentionally feeding images to non-corresponding modality-specific subnetworks of the surrogate model. Moreover, Multi Modality Collaborative Disruption strategy is devised to facilitate the attacker to comprehensively corrupt the informative content of person images by leveraging a multi modality feature collaborative metric disruption loss. Extensive experiments show that our MUA method can effectively attack the omni-modality re-id models, achieving 55.9%, 24.4%, 49.0% and 62.7% mean mAP Drop Rate, respectively., Comment: 9 pages,3 figures

Despite notable advancements in Retrieval-Augmented Generation (RAG) systems that expand large language model (LLM) capabilities through external retrieval, these systems often struggle to meet the complex and diverse needs of real-world industrial applications. The reliance on retrieval alone proves insufficient for extracting deep, domain-specific knowledge performing in logical reasoning from specialized corpora. To address this, we introduce sPecIalized KnowledgE and Rationale Augmentation Generation (PIKE-RAG), focusing on extracting, understanding, and applying specialized knowledge, while constructing coherent rationale to incrementally steer LLMs toward accurate responses. Recognizing the diverse challenges of industrial tasks, we introduce a new paradigm that classifies tasks based on their complexity in knowledge extraction and application, allowing for a systematic evaluation of RAG systems' problem-solving capabilities. This strategic approach offers a roadmap for the phased development and enhancement of RAG systems, tailored to meet the evolving demands of industrial applications. Furthermore, we propose knowledge atomizing and knowledge-aware task decomposition to effectively extract multifaceted knowledge from the data chunks and iteratively construct the rationale based on original query and the accumulated knowledge, respectively, showcasing exceptional performance across various benchmarks., Comment: 36 pages, 18 figures, technique report

The increasing availability of traffic videos functioning on a 24/7/365 time scale has the great potential of increasing the spatio-temporal coverage of traffic accidents, which will help improve traffic safety. However, analyzing footage from hundreds, if not thousands, of traffic cameras in a 24/7/365 working protocol remains an extremely challenging task, as current vision-based approaches primarily focus on extracting raw information, such as vehicle trajectories or individual object detection, but require laborious post-processing to derive actionable insights. We propose SeeUnsafe, a new framework that integrates Multimodal Large Language Model (MLLM) agents to transform video-based traffic accident analysis from a traditional extraction-then-explanation workflow to a more interactive, conversational approach. This shift significantly enhances processing throughput by automating complex tasks like video classification and visual grounding, while improving adaptability by enabling seamless adjustments to diverse traffic scenarios and user-defined queries. Our framework employs a severity-based aggregation strategy to handle videos of various lengths and a novel multimodal prompt to generate structured responses for review and evaluation and enable fine-grained visual grounding. We introduce IMS (Information Matching Score), a new MLLM-based metric for aligning structured responses with ground truth. We conduct extensive experiments on the Toyota Woven Traffic Safety dataset, demonstrating that SeeUnsafe effectively performs accident-aware video classification and visual grounding by leveraging off-the-shelf MLLMs. Source code will be available at \url{https://github.com/ai4ce/SeeUnsafe}.

We investigate the breaking of dark $SU(2)_d$ symmetry at different temperature scales, occurring after Peccei-Quinn symmetry breaking or following QCD symmetry breaking. We focus on assessing the potential of the hidden monopoles generated during this process to serve as dark matter candidate. Additionally, we examine the impact of axion-monopole interactions on the axion mass. When the phase transition occurs at extremely high temperature ($\sim 10^8 \mathrm{GeV}$), the contribution of monopoles to the axion mass through witten effect becomes non-negligible, playing a crucial role in accurately determining the axion relic density. Moreover, the stochastic gravitational wave background generated by dark phase transition and axionic domain wall collapse may offer a potential explanation for the low-frequency gravitational wave signals observed in PTA experiments.

Accurate estimation of fermionic observables is essential for advancing quantum physics and chemistry. The fermionic classical shadow (FCS) method offers an efficient framework for estimating these observables without requiring a transformation into a Pauli basis. However, the random matchgate circuits in FCS require polynomial-depth circuits with a brickwork structure, which presents significant challenges for near-term quantum devices with limited computational resources. To address this limitation, we introduce an adaptive-depth fermionic classical shadow (ADFCS) protocol designed to reduce the circuit depth while maintaining the estimation accuracy and the order of sample complexity. Through theoretical analysis and numerical fitting, we establish that the required depth for approximating a fermionic observable $H$ scales as $\max\{d^2_{\text{int}}(H)/\log n, d_{\text{int}}(H)\}$ where $d_{\text{int}}$ is the interaction distance of $H$. We demonstrate the effectiveness of the ADFCS protocol through numerical experiments, which show similar accuracy to the traditional FCS method while requiring significantly fewer resources. Additionally, we apply ADFCS to compute the expectation value of the Kitaev chain Hamiltonian, further validating its performance in practical scenarios. Our findings suggest that ADFCS can enable more efficient quantum simulations, reducing circuit depth while preserving the fidelity of quantum state estimations, offering a viable solution for near-term quantum devices., Comment: 28 pages, 9 figures

In this paper, we are interested in finding the global minimizer of a nonsmooth nonconvex unconstrained optimization problem. By combining the discrete consensus-based optimization (CBO) algorithm and the gradient descent method, we develop a novel CBO algorithm with an extra gradient descent scheme evaluated by the forward-difference technique on the function values, where only the objective function values are used in the proposed algorithm. First, we prove that the proposed algorithm can exhibit global consensus in an exponential rate in two senses and possess a unique global consensus point. Second, we evaluate the error estimate between the objective function value on the global consensus point and its global minimum. In particular, as the parameter $\beta$ tends to $\infty$, the error converges to zero and the convergence rate is $\mathcal{O}\left(\frac{\log\beta}{\beta}\right)$. Third, under some suitable assumptions on the objective function, we provide the number of iterations required for the mean square error in expectation to reach the desired accuracy. It is worth underlining that the theoretical analysis in this paper does not use the mean-field limit. Finally, we illustrate the improved efficiency and promising performance of our novel CBO method through some experiments on several nonconvex benchmark problems and the application to train deep neural networks.

For a class of sparse optimization problems with the penalty function of $\|(\cdot)_+\|_0$, we first characterize its local minimizers and then propose an extrapolated hard thresholding algorithm to solve such problems. We show that the iterates generated by the proposed algorithm with $\epsilon>0$ (where $\epsilon$ is the dry friction coefficient) have finite length, without relying on the Kurdyka-{\L}ojasiewicz inequality. Furthermore, we demonstrate that the algorithm converges to an $\epsilon$-local minimizer of this problem. For the special case that $\epsilon=0$, we establish that any accumulation point of the iterates is a local minimizer of the problem. Additionally, we analyze the convergence when an error term is present in the algorithm, showing that the algorithm still converges in the same manner as before, provided that the errors asymptotically approach zero. Finally, we conduct numerical experiments to verify the theoretical results of the proposed algorithm.

In this paper, we focus on finding the global minimizer of a general unconstrained nonsmooth nonconvex optimization problem. Taking advantage of the smoothing method and the consensus-based optimization (CBO) method, we propose a novel smoothing iterative consensus-based optimization (SICBO) algorithm. First, we prove that the solution process of the proposed algorithm here exponentially converges to a common stochastic consensus point almost surely. Second, we establish a detailed theoretical analysis to ensure the small enough error between the objective function value at the consensus point and the optimal function value, to the best of our knowledge, which provides the first theoretical guarantee to the global optimality of the proposed algorithm for nonconvex optimization problems. Moreover, unlike the previously introduced CBO methods, the theoretical results are valid for the cases that the objective function is nonsmooth, nonconvex and perhaps non-Lipschitz continuous. Finally, several numerical examples are performed to illustrate the effectiveness of our proposed algorithm for solving the global minimizer of the nonsmooth and nonconvex optimization problems.

Combinatorial medication recommendation(CMR) is a fundamental task of healthcare, which offers opportunities for clinical physicians to provide more precise prescriptions for patients with intricate health conditions, particularly in the scenarios of long-term medical care. Previous research efforts have sought to extract meaningful information from electronic health records (EHRs) to facilitate combinatorial medication recommendations. Existing learning-based approaches further consider the chemical structures of medications, but ignore the textual medication descriptions in which the functionalities are clearly described. Furthermore, the textual knowledge derived from the EHRs of patients remains largely underutilized. To address these issues, we introduce the Natural Language-Assisted Multi-modal Medication Recommendation(NLA-MMR), a multi-modal alignment framework designed to learn knowledge from the patient view and medication view jointly. Specifically, NLA-MMR formulates CMR as an alignment problem from patient and medication modalities. In this vein, we employ pretrained language models(PLMs) to extract in-domain knowledge regarding patients and medications, serving as the foundational representation for both modalities. In the medication modality, we exploit both chemical structures and textual descriptions to create medication representations. In the patient modality, we generate the patient representations based on textual descriptions of diagnosis, procedure, and symptom. Extensive experiments conducted on three publicly accessible datasets demonstrate that NLA-MMR achieves new state-of-the-art performance, with a notable average improvement of 4.72% in Jaccard score. Our source code is publicly available on https://github.com/jtan1102/NLA-MMR_CIKM_2024., Comment: 10 pages

The passive body-area electrostatic field has recently been aspiringly explored for wearable motion sensing, harnessing its two thrilling characteristics: full-body motion sensitivity and environmental sensitivity, which potentially empowers human activity recognition both independently and jointly from a single sensing front-end and theoretically brings significant competition against traditional inertial sensor that is incapable in environmental variations sensing. While most works focus on exploring the electrostatic field of a single body as the target, this work, for the first time, quantitatively evaluates the mutual effect of inter-body electrostatic fields and its contribution to collaborative activity recognition. A wearable electrostatic field sensing front-end and wrist-worn prototypes are built, and a sixteen-hour, manually annotated dataset is collected, involving an experiment of manipulating objects both independently and collaboratively. A regression model is finally used to recognize the collaborative activities among users. Despite the theoretical advantages of the body electrostatic field, the recognition of both single and collaborative activities shows unanticipated less-competitive recognition performance compared with the accelerometer. However, It is worth mentioning that this novel sensing modality improves the recognition F-score of user collaboration by 16\% in the fusion result of the two wearable motion sensing modalities, demonstrating the potential of bringing body electrostatic field as a complementary power-efficient signal for collaborative activity tracking using wearables.

Lately, a novel swarm intelligence model, namely the consensus-based optimization (CBO) algorithm, was introduced to deal with the global optimization problems. Limited by the conditions of Ito's formula, the convergence analysis of the previous CBO finite particle system mainly focuses on the problem with smooth objective function. With the help of smoothing method, this paper achieves a breakthrough by proposing an effective CBO algorithm for solving the global solution of a nonconvex, nonsmooth, and possible non-Lipschitz continuous minimization problem with theoretical analysis, which dose not rely on the mean-field limit. We indicate that the proposed algorithm exhibits a global consensus and converges to a common state with any initial data. Then, we give a more detailed error estimation on the objective function values along the state of the proposed algorithm towards the global minimum. Finally, some numerical examples are presented to illustrate the appreciable performance of the proposed method on solving the nonsmooth, nonconvex minimization problems.

Time series generation models are crucial for applications like data augmentation and privacy preservation. Most existing time series generation models are typically designed to generate data from one specified domain. While leveraging data from other domain for better generalization is proved to work in other application areas, this approach remains challenging for time series modeling due to the large divergence in patterns among different real world time series categories. In this paper, we propose a multi-domain time series diffusion model with domain prompts, named TimeDP. In TimeDP, we utilize a time series semantic prototype module which defines time series prototypes to represent time series basis, each prototype vector serving as "word" representing some elementary time series feature. A prototype assignment module is applied to extract the extract domain specific prototype weights, for learning domain prompts as generation condition. During sampling, we extract "domain prompt" with few-shot samples from the target domain and use the domain prompts as condition to generate time series samples. Experiments demonstrate that our method outperforms baselines to provide the state-of-the-art in-domain generation quality and strong unseen domain generation capability., Comment: AAAI 2025

Low-mass galaxies are the building blocks of massive galaxies in the framework of hierarchical structure formation. To enable detailed studies of galactic ecosystems in dwarf galaxies by spatially resolving different galactic components, we have carried out the Dwarf Galaxy Integral-field Survey (DGIS). This survey aims to acquire observations with spatial resolutions as high as 10 to 100 pc while maintaining reasonably high signal-to-noise ratios with VLT/MUSE and ANU-2.3m/WiFeS. The whole sample will be composed of 65 dwarf galaxies with $M_{\rm \ast}$ $<$ 10$^{9}$ $\rm M_{\odot}$, selected from the Spitzer Local Volume Legacy Survey. The overall scientific goals include studying baryonic cycles in dwarf galaxies, searching for off-nuclear (intermediate)-massive black holes, and quantifying the inner density profiles of dark matter. In this work, we describe the sample selection, data reduction, and high-level data products. By integrating the spectra over the field of view for each galaxy, we obtained the integrated gas-phase metallicity and discussed its dependence on stellar mass and SFR. We find that the overall relation between metallicity and stellar mass of our DGIS nearly follows the extrapolation from the higher mass end. Its dispersion does not decrease by invoking the dependence on SFR., Comment: 40 pages, 5 tables, and 24 figures. Submitted. DGIS website: https://www.dgisteam.com/index.html

4D video control is essential in video generation as it enables the use of sophisticated lens techniques, such as multi-camera shooting and dolly zoom, which are currently unsupported by existing methods. Training a video Diffusion Transformer (DiT) directly to control 4D content requires expensive multi-view videos. Inspired by Monocular Dynamic novel View Synthesis (MDVS) that optimizes a 4D representation and renders videos according to different 4D elements, such as camera pose and object motion editing, we bring pseudo 4D Gaussian fields to video generation. Specifically, we propose a novel framework that constructs a pseudo 4D Gaussian field with dense 3D point tracking and renders the Gaussian field for all video frames. Then we finetune a pretrained DiT to generate videos following the guidance of the rendered video, dubbed as GS-DiT. To boost the training of the GS-DiT, we also propose an efficient Dense 3D Point Tracking (D3D-PT) method for the pseudo 4D Gaussian field construction. Our D3D-PT outperforms SpatialTracker, the state-of-the-art sparse 3D point tracking method, in accuracy and accelerates the inference speed by two orders of magnitude. During the inference stage, GS-DiT can generate videos with the same dynamic content while adhering to different camera parameters, addressing a significant limitation of current video generation models. GS-DiT demonstrates strong generalization capabilities and extends the 4D controllability of Gaussian splatting to video generation beyond just camera poses. It supports advanced cinematic effects through the manipulation of the Gaussian field and camera intrinsics, making it a powerful tool for creative video production. Demos are available at https://wkbian.github.io/Projects/GS-DiT/., Comment: Project Page: https://wkbian.github.io/Projects/GS-DiT/

It is well-established that shear layer flows are linearly unstable provided the viscosity is small enough, when the horizontal Fourier number lies in some interval, between the so called lower and upper marginally stable curves. In this article, we prove that monotonic shear layers undergo a bifurcation near the upper marginally stable curve. In the particular case of the exponential boundary layer, we give evidences that this bifurcation is a Hopf bifurcation, which corresponds to the onset of moving rolls near the boundary. The onset of such rolls is the first step of one of the possible ``routes to chaos'' for shear layer flows.

Mamba is an efficient State Space Model (SSM) with linear computational complexity. Although SSMs are not suitable for handling non-causal data, Vision Mamba (ViM) methods still demonstrate good performance in tasks such as image classification and object detection. Recent studies have shown that there is a rich theoretical connection between state space models and attention variants. We propose a novel separable self attention method, for the first time introducing some excellent design concepts of Mamba into separable self-attention. To ensure a fair comparison with ViMs, we introduce VMINet, a simple yet powerful prototype architecture, constructed solely by stacking our novel attention modules with the most basic down-sampling layers. Notably, VMINet differs significantly from the conventional Transformer architecture. Our experiments demonstrate that VMINet has achieved competitive results on image classification and high-resolution dense prediction tasks.Code is available at: \url{https://github.com/yws-wxs/VMINet}.

Recent advancements in generative models have ignited substantial interest in dynamic 3D content creation (\ie, 4D generation). Existing approaches primarily rely on Score Distillation Sampling (SDS) to infer novel-view videos, typically leading to issues such as limited diversity, spatial-temporal inconsistency and poor prompt alignment, due to the inherent randomness of SDS. To tackle these problems, we propose AR4D, a novel paradigm for SDS-free 4D generation. Specifically, our paradigm consists of three stages. To begin with, for a monocular video that is either generated or captured, we first utilize pre-trained expert models to create a 3D representation of the first frame, which is further fine-tuned to serve as the canonical space. Subsequently, motivated by the fact that videos happen naturally in an autoregressive manner, we propose to generate each frame's 3D representation based on its previous frame's representation, as this autoregressive generation manner can facilitate more accurate geometry and motion estimation. Meanwhile, to prevent overfitting during this process, we introduce a progressive view sampling strategy, utilizing priors from pre-trained large-scale 3D reconstruction models. To avoid appearance drift introduced by autoregressive generation, we further incorporate a refinement stage based on a global deformation field and the geometry of each frame's 3D representation. Extensive experiments have demonstrated that AR4D can achieve state-of-the-art 4D generation without SDS, delivering greater diversity, improved spatial-temporal consistency, and better alignment with input prompts., Comment: TL;DR: We present a novel method for 4D generation from monocular videos without relying on SDS, delivering greater diversity, improved spatial-temporal consistency, and better alignment with input prompts. Project page: https://hanxinzhu-lab.github.io/AR4D/

Time series forecasting plays a crucial role in data mining, driving rapid advancements across numerous industries. With the emergence of large models, time series foundation models (TSFMs) have exhibited remarkable generalization capabilities, such as zero-shot learning, through large-scale pre-training. Meanwhile, Retrieval-Augmented Generation (RAG) methods have been widely employed to enhance the performance of foundation models on unseen data, allowing models to access to external knowledge. In this paper, we introduce TimeRAF, a Retrieval-Augmented Forecasting model that enhance zero-shot time series forecasting through retrieval-augmented techniques. We develop customized time series knowledge bases that are tailored to the specific forecasting tasks. TimeRAF employs an end-to-end learnable retriever to extract valuable information from the knowledge base. Additionally, we propose Channel Prompting for knowledge integration, which effectively extracts relevant information from the retrieved knowledge along the channel dimension. Extensive experiments demonstrate the effectiveness of our model, showing significant improvement across various domains and datasets.

Foundation models have revolutionized the paradigm of digital pathology, as they leverage general-purpose features to emulate real-world pathological practices, enabling the quantitative analysis of critical histological patterns and the dissection of cancer-specific signals. However, these static general features constrain the flexibility and pathological relevance in the ever-evolving needs of clinical applications, hindering the broad use of the current models. Here we introduce PathFiT, a dynamic feature learning method that can be effortlessly plugged into various pathology foundation models to unlock their adaptability. Meanwhile, PathFiT performs seamless implementation across diverse pathology applications regardless of downstream specificity. To validate PathFiT, we construct a digital pathology benchmark with over 20 terabytes of Internet and real-world data comprising 28 H\&E-stained tasks and 7 specialized imaging tasks including Masson's Trichrome staining and immunofluorescence images. By applying PathFiT to the representative pathology foundation models, we demonstrate state-of-the-art performance on 34 out of 35 tasks, with significant improvements on 23 tasks and outperforming by 10.20% on specialized imaging tasks. The superior performance and versatility of PathFiT open up new avenues in computational pathology., Comment: 49 pages, 14 figures

In this work, we investigate automatic design composition from multimodal graphic elements. Although recent studies have developed various generative models for graphic design, they usually face the following limitations: they only focus on certain subtasks and are far from achieving the design composition task; they do not consider the hierarchical information of graphic designs during the generation process. To tackle these issues, we introduce the layered design principle into Large Multimodal Models (LMMs) and propose a novel approach, called LaDeCo, to accomplish this challenging task. Specifically, LaDeCo first performs layer planning for a given element set, dividing the input elements into different semantic layers according to their contents. Based on the planning results, it subsequently predicts element attributes that control the design composition in a layer-wise manner, and includes the rendered image of previously generated layers into the context. With this insightful design, LaDeCo decomposes the difficult task into smaller manageable steps, making the generation process smoother and clearer. The experimental results demonstrate the effectiveness of LaDeCo in design composition. Furthermore, we show that LaDeCo enables some interesting applications in graphic design, such as resolution adjustment, element filling, design variation, etc. In addition, it even outperforms the specialized models in some design subtasks without any task-specific training., Comment: Project Page: $\href{https://elements2design.github.io/}{\text{elements2design}}$

Multi-party quantum steering is an important concept in quantum information theory and quantum mechanics, typically related to quantum entanglement and quantum nonlocality. It enables precise manipulation of large quantum systems, which is essential for large-scale quantum computing, simulations, and quantum communication. Recently, a quantum steering certification for any three-qubit generalized Greenberger-Horne-Zeilinger (GGHZ) states based on the fine-grained steering inequality was proved [Quantum Studies: Mathematics and Foundations, 2022, 9(2): 175-198]. Here we provide an experimental proposal to prepare the GGHZ states in photon system. The measurement observalbes in each party can be realized by different polarization optical elements. By choosing the angles of the waveplates, our experiment proposal can observe the maximum quantum violation for any three-qubit GGHZ states. Our proposal can be easily extended to high-dimensional qubits and multi-photon GHZ states, which provides a method to study the complex multi-party quantum protocols.

In recent years, Artificial Intelligence Weather Prediction (AIWP) models have achieved performance comparable to, or even surpassing, traditional Numerical Weather Prediction (NWP) models by leveraging reanalysis data. However, a less-explored approach involves training AIWP models directly on observational data, enhancing computational efficiency and improving forecast accuracy by reducing the uncertainties introduced through data assimilation processes. In this study, we propose OMG-HD, a novel AI-based regional high-resolution weather forecasting model designed to make predictions directly from observational data sources, including surface stations, radar, and satellite, thereby removing the need for operational data assimilation. Our evaluation shows that OMG-HD outperforms both the European Centre for Medium-Range Weather Forecasts (ECMWF)'s high-resolution operational forecasting system, IFS-HRES, and the High-Resolution Rapid Refresh (HRRR) model at lead times of up to 12 hours across the contiguous United States (CONUS) region. We achieve up to a 13% improvement on RMSE for 2-meter temperature, 17% on 10-meter wind speed, 48% on 2-meter specific humidity, and 32% on surface pressure compared to HRRR. Our method shows that it is possible to use AI-driven approaches for rapid weather predictions without relying on NWP-derived weather fields as model input. This is a promising step towards using observational data directly to make operational forecasts with AIWP models.

Transportation equity research has traditionally emphasized service accessibility and destination reachability while often overlooking the critical aspects of service quality, such as infrequent schedules or overcrowded vehicles. This oversight can lead to a skewed understanding of equity, as high accessibility does not guarantee high-quality service. Addressing this gap, we propose a transportation equity index called the multi-dimensional, high-granularity (MDHG) index. Such an index considers service accessibility and quality alongside population demographics. This approach ensures that areas with high accessibility but low service quality are recognized as inequitable. The MDHG Index addresses service performance by incorporating performance data with temporal variations based on actual trip data, thus offering a more nuanced view of transportation equity that reflects the real-world experiences of service users. Furthermore, to effectively identify and address the needs at the user level, we need to use a highly granular population dataset. Due to the low granularity of census and other open-source datasets, we opted to use a highly granular synthetic dataset. To test out the MDHG Index, we coupled a highly granular synthetic population dataset with data from NYC Citi Bike expansion to use as a case study to assess changes in accessibility and service quality before and after the expansion. The MDHG approach effectively identified areas that improved post-expansion and highlighted those requiring further enhancement, thus showing the effectiveness of the index in targeted improvements for transportation equity.

Introduction. Generative Artificial Intelligence, particularly large language models (LLMs), offers transformative potential for Health Economics and Outcomes Research (HEOR). However, evaluating the quality, transparency, and rigor of LLM-assisted research lacks standardized guidance. This article introduces the ELEVATE AI LLMs framework and checklist, designed to support researchers and reviewers in assessing LLM use in HEOR. Methods. The ELEVATE AI LLMs framework was developed through a targeted review of existing guidelines and evaluation frameworks. The framework comprises ten evaluation domains, including model characteristics, accuracy, comprehensiveness, and fairness. The accompanying checklist operationalizes the framework. To validate the framework, we applied it to two published studies, demonstrating its usability across different HEOR tasks. Results. The ELEVATE AI LLMs framework provides a comprehensive structure for evaluating LLM-assisted research, while the checklist facilitates practical application. Validation of the framework and checklist on studies of systematic literature reviews and health economic modeling highlighted their ability to identify strengths and gaps in reporting. Limitations. While the ELEVATE AI LLMs framework provides robust guidance, its broader generalizability and applicability to diverse HEOR tasks require further empirical testing. Additionally, several metrics adapted from computer science need further validation in HEOR contexts. Conclusion. The ELEVATE AI LLMs framework and checklist fill a critical gap in HEOR by offering structured guidance for evaluating LLM-assisted research. By promoting transparency, accuracy, and reproducibility, they aim to standardize and improve the integration of LLMs into HEOR, ensuring their outputs meet the field's rigorous standards., Comment: 4 Tables, 1 Figure, Supplemental Material

We introduce ChatGarment, a novel approach that leverages large vision-language models (VLMs) to automate the estimation, generation, and editing of 3D garments from images or text descriptions. Unlike previous methods that struggle in real-world scenarios or lack interactive editing capabilities, ChatGarment can estimate sewing patterns from in-the-wild images or sketches, generate them from text descriptions, and edit garments based on user instructions, all within an interactive dialogue. These sewing patterns can then be draped into 3D garments, which are easily animatable and simulatable. This is achieved by finetuning a VLM to directly generate a JSON file that includes both textual descriptions of garment types and styles, as well as continuous numerical attributes. This JSON file is then used to create sewing patterns through a programming parametric model. To support this, we refine the existing programming model, GarmentCode, by expanding its garment type coverage and simplifying its structure for efficient VLM fine-tuning. Additionally, we construct a large-scale dataset of image-to-sewing-pattern and text-to-sewing-pattern pairs through an automated data pipeline. Extensive evaluations demonstrate ChatGarment's ability to accurately reconstruct, generate, and edit garments from multimodal inputs, highlighting its potential to revolutionize workflows in fashion and gaming applications. Code and data will be available at https://chatgarment.github.io/.

Recent advancements in video autoencoders (Video AEs) have significantly improved the quality and efficiency of video generation. In this paper, we propose a novel and compact video autoencoder, VidTwin, that decouples video into two distinct latent spaces: Structure latent vectors, which capture overall content and global movement, and Dynamics latent vectors, which represent fine-grained details and rapid movements. Specifically, our approach leverages an Encoder-Decoder backbone, augmented with two submodules for extracting these latent spaces, respectively. The first submodule employs a Q-Former to extract low-frequency motion trends, followed by downsampling blocks to remove redundant content details. The second averages the latent vectors along the spatial dimension to capture rapid motion. Extensive experiments show that VidTwin achieves a high compression rate of 0.20% with high reconstruction quality (PSNR of 28.14 on the MCL-JCV dataset), and performs efficiently and effectively in downstream generative tasks. Moreover, our model demonstrates explainability and scalability, paving the way for future research in video latent representation and generation. Our code has been released at https://github.com/microsoft/VidTok/tree/main/vidtwin.

The item cold-start problem is crucial for online recommender systems, as the success of the cold-start phase determines whether items can transition into popular ones. Prompt learning, a powerful technique used in natural language processing (NLP) to address zero- or few-shot problems, has been adapted for recommender systems to tackle similar challenges. However, existing methods typically rely on content-based properties or text descriptions for prompting, which we argue may be suboptimal for cold-start recommendations due to 1) semantic gaps with recommender tasks, 2) model bias caused by warm-up items contribute most of the positive feedback to the model, which is the core of the cold-start problem that hinders the recommender quality on cold-start items. We propose to leverage high-value positive feedback, termed pinnacle feedback as prompt information, to simultaneously resolve the above two problems. We experimentally prove that compared to the content description proposed in existing works, the positive feedback is more suitable to serve as prompt information by bridging the semantic gaps. Besides, we propose item-wise personalized prompt networks to encode pinnaclce feedback to relieve the model bias by the positive feedback dominance problem. Extensive experiments on four real-world datasets demonstrate the superiority of our model over state-of-the-art methods. Moreover, PROMO has been successfully deployed on a popular short-video sharing platform, a billion-user scale commercial short-video application, achieving remarkable performance gains across various commercial metrics within cold-start scenarios

Spatially varying regularization accommodates the deformation variations that may be necessary for different anatomical regions during deformable image registration. Historically, optimization-based registration models have harnessed spatially varying regularization to address anatomical subtleties. However, most modern deep learning-based models tend to gravitate towards spatially invariant regularization, wherein a homogenous regularization strength is applied across the entire image, potentially disregarding localized variations. In this paper, we propose a hierarchical probabilistic model that integrates a prior distribution on the deformation regularization strength, enabling the end-to-end learning of a spatially varying deformation regularizer directly from the data. The proposed method is straightforward to implement and easily integrates with various registration network architectures. Additionally, automatic tuning of hyperparameters is achieved through Bayesian optimization, allowing efficient identification of optimal hyperparameters for any given registration task. Comprehensive evaluations on publicly available datasets demonstrate that the proposed method significantly improves registration performance and enhances the interpretability of deep learning-based registration, all while maintaining smooth deformations., Comment: Code available at http://bit.ly/3BrXGxz

It is clinically crucial and potentially very beneficial to be able to analyze and model directly the spatial distributions of cells in histopathology whole slide images (WSI). However, most existing WSI datasets lack cell-level annotations, owing to the extremely high cost over giga-pixel images. Thus, it remains an open question whether deep learning models can directly and effectively analyze WSIs from the semantic aspect of cell distributions. In this work, we construct a large-scale WSI dataset with more than 5 billion cell-level annotations, termed WSI-Cell5B, and a novel hierarchical Cell Cloud Transformer (CCFormer) to tackle these challenges. WSI-Cell5B is based on 6,998 WSIs of 11 cancers from The Cancer Genome Atlas Program, and all WSIs are annotated per cell by coordinates and types. To the best of our knowledge, WSI-Cell5B is the first WSI-level large-scale dataset integrating cell-level annotations. On the other hand, CCFormer formulates the collection of cells in each WSI as a cell cloud and models cell spatial distribution. Specifically, Neighboring Information Embedding (NIE) is proposed to characterize the distribution of cells within the neighborhood of each cell, and a novel Hierarchical Spatial Perception (HSP) module is proposed to learn the spatial relationship among cells in a bottom-up manner. The clinical analysis indicates that WSI-Cell5B can be used to design clinical evaluation metrics based on counting cells that effectively assess the survival risk of patients. Extensive experiments on survival prediction and cancer staging show that learning from cell spatial distribution alone can already achieve state-of-the-art (SOTA) performance, i.e., CCFormer strongly outperforms other competing methods.

Histopathology plays a critical role in medical diagnostics, with whole slide images (WSIs) offering valuable insights that directly influence clinical decision-making. However, the large size and complexity of WSIs may pose significant challenges for deep learning models, in both computational efficiency and effective representation learning. In this work, we introduce Pixel-Mamba, a novel deep learning architecture designed to efficiently handle gigapixel WSIs. Pixel-Mamba leverages the Mamba module, a state-space model (SSM) with linear memory complexity, and incorporates local inductive biases through progressively expanding tokens, akin to convolutional neural networks. This enables Pixel-Mamba to hierarchically combine both local and global information while efficiently addressing computational challenges. Remarkably, Pixel-Mamba achieves or even surpasses the quantitative performance of state-of-the-art (SOTA) foundation models that were pretrained on millions of WSIs or WSI-text pairs, in a range of tumor staging and survival analysis tasks, {\bf even without requiring any pathology-specific pretraining}. Extensive experiments demonstrate the efficacy of Pixel-Mamba as a powerful and efficient framework for end-to-end WSI analysis.

For efficient human-agent interaction, an agent should proactively recognize their target user and prepare for upcoming interactions. We formulate this challenging problem as the novel task of jointly forecasting a person's intent to interact with the agent, their attitude towards the agent and the action they will perform, from the agent's (egocentric) perspective. So we propose \emph{SocialEgoNet} - a graph-based spatiotemporal framework that exploits task dependencies through a hierarchical multitask learning approach. SocialEgoNet uses whole-body skeletons (keypoints from face, hands and body) extracted from only 1 second of video input for high inference speed. For evaluation, we augment an existing egocentric human-agent interaction dataset with new class labels and bounding box annotations. Extensive experiments on this augmented dataset, named JPL-Social, demonstrate \emph{real-time} inference and superior performance (average accuracy across all tasks: 83.15\%) of our model outperforming several competitive baselines. The additional annotations and code will be available upon acceptance.

Knowledge stored in large language models requires timely updates to reflect the dynamic nature of real-world information. To update the knowledge, most knowledge editing methods focus on the low layers, since recent probes into the knowledge recall process reveal that the answer information is enriched in low layers. However, these probes only and could only reveal critical recall stages for the original answers, while the goal of editing is to rectify model's prediction for the target answers. This inconsistency indicates that both the probe approaches and the associated editing methods are deficient. To mitigate the inconsistency and identify critical editing regions, we propose a contrast-based probe approach, and locate two crucial stages where the model behavior diverges between the original and target answers: Information Enrichment in low layers and Probability Promotion in high layers. Building upon the insights, we develop the Joint knowledge Editing for information Enrichment and probability Promotion (JEEP) method, which jointly edits both the low and high layers to modify the two critical recall stages. Considering the mutual interference and growing forgetting due to dual modifications, JEEP is designed to ensure that updates to distinct regions share the same objectives and are complementary. We rigorously evaluate JEEP by editing up to thousands of facts on various models, i.e., GPT-J (6B) and LLaMA (7B), and addressing diverse editing objectives, i.e., adding factual and counterfactual knowledge. In all tested scenarios, JEEP achieves best performances, validating the effectiveness of the revealings of our probe approach and the designs of our editing method. Our code and data are available at https://github.com/Eric8932/JEEP.

Early-warning signals of delicate design are always used to predict critical transitions in complex systems, which makes it possible to render the systems far away from the catastrophic state by introducing timely interventions. Traditional signals including the dynamical network biomarker (DNB), based on statistical properties such as variance and autocorrelation of nodal dynamics, overlook directional interactions and thus have limitations in capturing underlying mechanisms and simultaneously sustaining robustness against noise perturbations. This paper therefore introduces a framework of causal network markers (CNMs) by incorporating causality indicators, which reflect the directional influence between variables. Actually, to detect and identify the tipping points ahead of critical transition, two markers are designed: CNM-GC for linear causality and CNM-TE for non-linear causality, as well as a functional representation of different causality indicators and a clustering technique to verify the system's dominant group. Through demonstrations using benchmark models and real-world datasets of epileptic seizure, the framework of CNMs shows higher predictive power and accuracy than the traditional DNB indicator. It is believed that, due to the versatility and scalability, the CNMs are suitable for comprehensively evaluating the systems. The most possible direction for application includes the identification of tipping points in clinical disease., Comment: 16 pages, 4 figures

Nanoconfined water plays an indispensable role in various phenomena in biology, chemistry, and engineering. It exhibits many abnormal properties compared to bulk water, especially under strong confinement. However, the origin of those anomalies is still elusive due to the lack of structural information on hydrogen-bonding networks. Considering the inhomogeneity of the nanocavity and the tiny amount of water molecules, conventional optical spectroscopies and nuclear magnetic resonance (NMR) fail to realize the structure analysis of nanoconfined water. Here, we addressed this issue by combining scanning probe microscopy (SPM) with advanced quantum sensing(QS) based on an atomic-size quantum sensor like nitrogen-vacancy (NV) center in diamond, which can apply the nanoscale-NMR for characterizing both the dynamics and structure of confined water at ambient conditions. We built a two-dimensional (2D) nanoconfined water system with a hexagonal-boron nitride (hBN) flake and a hydrophilic diamond surface. By using the SPM tip to measure the confinement size precisely, we observed a critical confinement size of ~2 nm, below which the water diffusion was significantly suppressed and the hydrogen-bonding network of water showed an ordered structure. Meanwhile, molecular dynamics (MD) simulation revealed a solid-like water contact layer on the diamond surface under strong confinement, which also reproduced the measured nanoscale-NMR spectra and confirmed the liquid-solid phase transition observed in the experiments. Notably, with this new SPM-QS platform, our results showed a promising way to elucidate the abnormal properties of nanoconfined water in future applications.

Analog circuits are crucial in modern electronic systems, and automating their design has attracted significant research interest. One of major challenges is topology synthesis, which determines circuit components and their connections. Recent studies explore large language models (LLM) for topology synthesis. However, the scenarios addressed by these studies do not align well with practical applications. Specifically, existing work uses vague design requirements as input and outputs an ideal model, but detailed structural requirements and device-level models are more practical. Moreover, current approaches either formulate topology synthesis as graph generation or Python code generation, whereas practical topology design is a complex process that demands extensive design knowledge. In this work, we propose AnalogXpert, a LLM-based agent aiming at solving practical topology synthesis problem by incorporating circuit design expertise into LLMs. First, we represent analog topology as SPICE code and introduce a subcircuit library to reduce the design space, in the same manner as experienced designers. Second, we decompose the problem into two sub-task (i.e., block selection and block connection) through the use of CoT and incontext learning techniques, to mimic the practical design process. Third, we introduce a proofreading strategy that allows LLMs to incrementally correct the errors in the initial design, akin to human designers who iteratively check and adjust the initial topology design to ensure accuracy. Finally, we construct a high-quality benchmark containing both real data (30) and synthetic data (2k). AnalogXpert achieves 40% and 23% success rates on the synthetic dataset and real dataset respectively, which is markedly better than those of GPT-4o (3% on both the synthetic dataset and the real dataset).

We consider a Keller-Segel model with non-linear porous medium type diffusion and non-local attractive power law interaction, focusing on potentials that are less singular than Newtonian interaction. Here, the nonlinear diffusion is chosen to be $m=2-\frac{2s}{d}$, in which case the steady states are compactly supported. We analyse under what conditions on the initial data the regime that attractive forces are stronger than diffusion occurs and classify the conditions for global existence and finite time blow-up of solutions. It is shown that there exists a threshold value which is characterized by the optimal constant of a variant of the Hardy-Littlewood-Sobolev inequality. Specifically, the solution will exist globally if the initial data is below the threshold, while the solution blows up in finite time when the initial data is above the threshold.

Recent advancements in large language models (LLMs) have shown impressive versatility across various tasks. To eliminate its hallucinations, retrieval-augmented generation (RAG) has emerged as a powerful approach, leveraging external knowledge sources like knowledge graphs (KGs). In this paper, we study the task of KG-driven RAG and propose a novel Similar Graph Enhanced Retrieval-Augmented Generation (SimGRAG) method. It effectively addresses the challenge of aligning query texts and KG structures through a two-stage process: (1) query-to-pattern, which uses an LLM to transform queries into a desired graph pattern, and (2) pattern-to-subgraph, which quantifies the alignment between the pattern and candidate subgraphs using a graph semantic distance (GSD) metric. We also develop an optimized retrieval algorithm that efficiently identifies the top-$k$ subgraphs within 1-second latency on a 10-million-scale KG. Extensive experiments show that SimGRAG outperforms state-of-the-art KG-driven RAG methods in both question answering and fact verification, offering superior plug-and-play usability and scalability.

Based on a determinantal formula for the higher derivative of a quotient of two functions, we first present the determinantal expressions of Eulerian polynomials and Andre polynomials. In particular, we discover that the Euler number (number of alternating permutations) can be expressed as a lower Hessenberg determinant. We then investigate the determinantal representations of the up-down run polynomials and the types A and B alternating run polynomials. As applications, we deduce several new recurrence relations, which imply the multiplicity of -1 in these three kinds of polynomials. And then, we provide two determinantal representations for the alternating run polynomials of dual Stirling permutations. In particular, we discover a close connection between the alternating run polynomials of dual Stirling permutations and the type B Eulerian polynomials., Comment: 26 pages

Encoding video content into compact latent tokens has become a fundamental step in video generation and understanding, driven by the need to address the inherent redundancy in pixel-level representations. Consequently, there is a growing demand for high-performance, open-source video tokenizers as video-centric research gains prominence. We introduce VidTok, a versatile video tokenizer that delivers state-of-the-art performance in both continuous and discrete tokenizations. VidTok incorporates several key advancements over existing approaches: 1) model architecture such as convolutional layers and up/downsampling modules; 2) to address the training instability and codebook collapse commonly associated with conventional Vector Quantization (VQ), we integrate Finite Scalar Quantization (FSQ) into discrete video tokenization; 3) improved training strategies, including a two-stage training process and the use of reduced frame rates. By integrating these advancements, VidTok achieves substantial improvements over existing methods, demonstrating superior performance across multiple metrics, including PSNR, SSIM, LPIPS, and FVD, under standardized evaluation settings., Comment: Code & Models: https://github.com/microsoft/VidTok

This study investigates the trade-offs between fairness, privacy, and utility in image classification using machine learning (ML). Recent research suggests that generalization techniques can improve the balance between privacy and utility. One focus of this work is sharpness-aware training (SAT) and its integration with differential privacy (DP-SAT) to further improve this balance. Additionally, we examine fairness in both private and non-private learning models trained on datasets with synthetic and real-world biases. We also measure the privacy risks involved in these scenarios by performing membership inference attacks (MIAs) and explore the consequences of eliminating high-privacy risk samples, termed outliers. Moreover, we introduce a new metric, named \emph{harmonic score}, which combines accuracy, privacy, and fairness into a single measure. Through empirical analysis using generalization techniques, we achieve an accuracy of 81.11\% under $(8, 10^{-5})$-DP on CIFAR-10, surpassing the 79.5\% reported by De et al. (2022). Moreover, our experiments show that memorization of training samples can begin before the overfitting point, and generalization techniques do not guarantee the prevention of this memorization. Our analysis of synthetic biases shows that generalization techniques can amplify model bias in both private and non-private models. Additionally, our results indicate that increased bias in training data leads to reduced accuracy, greater vulnerability to privacy attacks, and higher model bias. We validate these findings with the CelebA dataset, demonstrating that similar trends persist with real-world attribute imbalances. Finally, our experiments show that removing outlier data decreases accuracy and further amplifies model bias., Comment: Published as a conference paper at the 25th Privacy Enhancing Technologies Symposium (PETS 2025)

Image editing has advanced significantly with the development of diffusion models using both inversion-based and instruction-based methods. However, current inversion-based approaches struggle with big modifications (e.g., adding or removing objects) due to the structured nature of inversion noise, which hinders substantial changes. Meanwhile, instruction-based methods often constrain users to black-box operations, limiting direct interaction for specifying editing regions and intensity. To address these limitations, we propose BrushEdit, a novel inpainting-based instruction-guided image editing paradigm, which leverages multimodal large language models (MLLMs) and image inpainting models to enable autonomous, user-friendly, and interactive free-form instruction editing. Specifically, we devise a system enabling free-form instruction editing by integrating MLLMs and a dual-branch image inpainting model in an agent-cooperative framework to perform editing category classification, main object identification, mask acquisition, and editing area inpainting. Extensive experiments show that our framework effectively combines MLLMs and inpainting models, achieving superior performance across seven metrics including mask region preservation and editing effect coherence., Comment: WebPage available at https://liyaowei-stu.github.io/project/BrushEdit/