搜索结果：Dive

Towards human-level visual understanding, visual commonsense generation has been introduced to generate commonsense inferences beyond images. However, current research on visual commonsense generation has overlooked an important human cognitive ability: generating descriptive and diverse inferences. In this work, we propose a novel visual commonsense generation framework, called DIVE, which aims to improve the descriptiveness and diversity of generated inferences. DIVE involves two methods, generic inference filtering and contrastive retrieval learning, which address the limitations of existing visual commonsense resources and training objectives. Experimental results verify that DIVE outperforms state-of-the-art models for visual commonsense generation in terms of both descriptiveness and diversity, while showing a superior quality in generating unique and novel inferences. Notably, DIVE achieves human-level descriptiveness and diversity on Visual Commonsense Graphs. Furthermore, human evaluations confirm that DIVE aligns closely with human judgments on descriptiveness and diversity\footnote{Our code and dataset are available at https://github.com/Park-ing-lot/DIVE.

Recent advances in large language models (LLMs) have demonstrated the effectiveness of Iterative Self-Improvement (ISI) techniques. However, continuous training on self-generated data leads to reduced output diversity, a limitation particularly critical in reasoning tasks where diverse solution paths are essential. We present DIVE (Diversified Iterative Self-Improvement), a novel framework that addresses this challenge through two key components: Sample Pool Expansion for broader solution exploration, and Data Selection for balancing diversity and quality in preference pairs. Experiments on MATH and GSM8k datasets show that DIVE achieves a 10% to 45% relative increase in output diversity metrics while maintaining performance quality compared to vanilla ISI. Our ablation studies confirm both components' significance in achieving these improvements. Code is available at https://github.com/qinyiwei/DIVE.

Large language models (LLMs) with the Mixture-of-Experts (MoE) architecture achieve high cost-efficiency by selectively activating a subset of the parameters. Despite the inference efficiency of MoE LLMs, the training of extensive experts from scratch incurs substantial overhead, whereas reconstructing a dense LLM into an MoE LLM significantly reduces the training budget. However, existing reconstruction methods often overlook the diversity among experts, leading to potential redundancy. In this paper, we come up with the observation that a specific LLM exhibits notable diversity after being pruned on different calibration datasets, based on which we present a Diversity-Enhanced reconstruction method named DIVE. The recipe of DIVE includes domain affinity mining, pruning-based expert reconstruction, and efficient retraining. Specifically, the reconstruction includes pruning and reassembly of the feed-forward network (FFN) module. After reconstruction, we efficiently retrain the model on routers, experts and normalization modules. We implement DIVE on Llama-style LLMs with open-source training corpora. Experiments show that DIVE achieves training efficiency with minimal accuracy tra

Evaluating Retrieval-Augmented Generation (RAG) systems using static multi-turn datasets fails to capture the dynamic nature of real-world dialogues. Existing evaluation methods rely on predefined datasets, which restrict them to static, one-directional queries and limit their ability to capture the adaptive, context-dependent performance of RAG systems in interactive, multi-turn settings. Thus, we introduce the RAG-DIVE, a Dynamic Interactive Validation and Evaluation approach, that simulates user interactions with RAG systems. RAG-DIVE leverages an LLM to generate multi-turn conversations dynamically and is organized into three components. The dialogue generation stage consists of the (1) Conversation Generator, which simulates a user by creating multi-turn queries, and the (2) Conversation Validator, which filters and corrects invalid or low-quality outputs to ensure coherent conversations. The evaluation stage is handled by the (3) Conversation Evaluator, which assesses the RAG system's performance across the entire dialogue and generates both per-turn and multi-turn metrics that provide an aggregated view of system behavior. We validated RAG-DIVE through two experimental setup

Building on the success of diffusion models in image generation and editing, video editing has recently gained substantial attention. However, maintaining temporal consistency and motion alignment still remains challenging. To address these issues, this paper proposes DINO-guided Video Editing (DIVE), a framework designed to facilitate subject-driven editing in source videos conditioned on either target text prompts or reference images with specific identities. The core of DIVE lies in leveraging the powerful semantic features extracted from a pretrained DINOv2 model as implicit correspondences to guide the editing process. Specifically, to ensure temporal motion consistency, DIVE employs DINO features to align with the motion trajectory of the source video. For precise subject editing, DIVE incorporates the DINO features of reference images into a pretrained text-to-image model to learn Low-Rank Adaptations (LoRAs), effectively registering the target subject's identity. Extensive experiments on diverse real-world videos demonstrate that our framework can achieve high-quality editing results with robust motion consistency, highlighting the potential of DINO to contribute to video e

Person re-identification (Re-ID) often faces challenges due to variations in human poses and camera viewpoints, which significantly affect the appearance of individuals across images. Existing datasets frequently lack diversity and scalability in these aspects, hindering the generalization of Re-ID models to new camera systems or environments. To overcome this, we propose Pose-dIVE, a novel data augmentation approach that incorporates sparse and underrepresented human pose and camera viewpoint examples into the training data, addressing the limited diversity in the original training data distribution. Our objective is to augment the training dataset to enable existing Re-ID models to learn features unbiased by human pose and camera viewpoint variations. By conditioning the diffusion model on both the human pose and camera viewpoint through the SMPL model, our framework generates augmented training data with diverse human poses and camera viewpoints. Experimental results demonstrate the effectiveness of our method in addressing human pose bias and enhancing the generalizability of Re-ID models compared to other data augmentation-based Re-ID approaches.

Data-driven artificial intelligence (AI) approaches are fundamentally transforming the discovery of new materials. Despite the unprecedented availability of materials data in the scientific literature, much of this information remains trapped in unstructured figures and tables, hindering the construction of large language model (LLM)-based AI agent for automated materials design. Here, we present the Descriptive Interpretation of Visual Expression (DIVE) multi-agent workflow, which systematically reads and organizes experimental data from graphical elements in scientific literatures. We focus on solid-state hydrogen storage materials-a class of materials central to future clean-energy technologies and demonstrate that DIVE markedly improves the accuracy and coverage of data extraction compared to the direct extraction by multimodal models, with gains of 10-15% over commercial models and over 30% relative to open-source models. Building on a curated database of over 30,000 entries from 4,000 publications, we establish a rapid inverse design workflow capable of identifying previously unreported hydrogen storage compositions in two minutes. The proposed AI workflow and agent design ar

Collecting multi-view driving scenario videos to enhance the performance of 3D visual perception tasks presents significant challenges and incurs substantial costs, making generative models for realistic data an appealing alternative. Yet, the videos generated by recent works suffer from poor quality and spatiotemporal consistency, undermining their utility in advancing perception tasks under driving scenarios. To address this gap, we propose DiVE, a diffusion transformer-based generative framework meticulously engineered to produce high-fidelity, temporally coherent, and cross-view consistent multi-view videos, aligning seamlessly with bird's-eye view layouts and textual descriptions. DiVE leverages a unified cross-attention and a SketchFormer to exert precise control over multimodal data, while incorporating a view-inflated attention mechanism that adds no extra parameters, thereby guaranteeing consistency across views. Despite these advancements, synthesizing high-resolution videos under multimodal constraints introduces dual challenges: investigating the optimal classifier-free guidance coniguration under intricate multi-condition inputs and mitigating excessive computational l

Claude Code is an agentic coding tool that can run shell commands, edit files, and call external services on behalf of the user. This study describes its comprehensive architecture by analyzing the publicly available TypeScript source code and further comparing it with OpenClaw, an independent open-source AI agent system that answers many of the same design questions from a different deployment context. Our analysis identifies five human values, philosophies, and needs that motivate the architecture (human decision authority, safety and security, reliable execution, capability amplification, and contextual adaptability) and traces them through thirteen design principles to specific implementation choices. The core of the system is a simple while-loop that calls the model, runs tools, and repeats. Most of the code, however, lives in the systems around this loop: a permission system with seven modes and an ML-based classifier, a five-layer compaction pipeline for context management, four extensibility mechanisms (MCP, plugins, skills, and hooks), a subagent delegation mechanism with worktree isolation, and append-oriented session storage. A comparison with OpenClaw, a multi-channel p

In this report, we present the winning solution that achieved the 1st place in the Complex Video Reasoning & Robustness Evaluation Challenge 2025. This challenge evaluates the ability to generate accurate natural language answers to questions about diverse, real-world video clips. It uses the Complex Video Reasoning and Robustness Evaluation Suite (CVRR-ES) benchmark, which consists of 214 unique videos and 2,400 question-answer pairs spanning 11 categories. Our method, DIVE (Deep-search Iterative Video Exploration), adopts an iterative reasoning approach, in which each input question is semantically decomposed and solved through stepwise reasoning and progressive inference. This enables our system to provide highly accurate and contextually appropriate answers to even the most complex queries. Applied to the CVRR-ES benchmark, our approach achieves 81.44% accuracy on the test set, securing the top position among all participants. This report details our methodology and provides a comprehensive analysis of the experimental results, demonstrating the effectiveness of our iterative reasoning framework in achieving robust video question answering. The code is available at https://

Primal heuristics are important for solving mixed integer linear programs, because they find feasible solutions that facilitate branch and bound search. A prominent group of primal heuristics are diving heuristics. They iteratively modify and resolve linear programs to conduct a depth-first search from any node in the search tree. Existing divers rely on generic decision rules that fail to exploit structural commonality between similar problem instances that often arise in practice. Therefore, we propose L2Dive to learn specific diving heuristics with graph neural networks: We train generative models to predict variable assignments and leverage the duality of linear programs to make diving decisions based on the model's predictions. L2Dive is fully integrated into the open-source solver SCIP. We find that L2Dive outperforms standard divers to find better feasible solutions on a range of combinatorial optimization problems. For real-world applications from server load balancing and neural network verification, L2Dive improves the primal-dual integral by up to 7% (35%) on average over a tuned (default) solver baseline and reduces average solving time by 20% (29%).

Large Vision Language Models (LVLMs) possess extensive text knowledge but struggles to utilize this knowledge for fine-grained image recognition, often failing to differentiate between visually similar categories. Existing fine-tuning methods using Reinforcement Learning (RL) with exact-match reward signals are often brittle, encourage memorization of training categories, and fail to elicit differential reasoning needed for generalization to unseen classes. To address this, we propose $\textbf{DiVE-k}$, $\textbf{Di}$fferential $\textbf{V}$isual r$\textbf{E}$asoning using top-$\textbf{k}$ generations, framework that leverages model's own top-k predictions as a training signal. For each training image, DiVE-k creates a multiple-choice question from the model's top-k outputs and uses RL to train the model to select the correct answer. This approach requires the model to perform fine-grained differential reasoning among plausible options and provides a simple, verifiable reward signal that mitigates memorization and improves generalization. Experiments on five standard fine-grained datasets show that our method significantly outperforms existing approaches. In the standard base-to-nove

Low Earth Orbit (LEO) satellite networks are an important part of the global communication infrastructure today. Despite ongoing efforts to improve their resilience, they remain vulnerable to component damage and deorbiting under harsh space weather conditions. Prior work identified a modest but noticeable impact on LEO satellite network performance during solar storms, typically manifesting as an immediate rise in packet loss and a sustained increase in round-trip time (RTT). However, these studies offer only coarse-grained insights and do not capture the nuanced spatial and temporal patterns of disruption across the LEO network. In this paper, we conduct a deep dive into the impact of solar storms on LEO satellite communications. By localizing the impact of increased atmospheric drag at the level of individual satellites and orbits, we reveal significant heterogeneity in how different parts of the network are affected. We find that the degree of performance degradation varies significantly across geographic regions, depending on satellite positioning during the storm. Specifically, we find that (i) not all satellite orbits are equally vulnerable, (ii) within a given orbit, certai

Large language models (LLMs) have shown impressive effectiveness in various software engineering tasks, including automated program repair (APR). In this study, we take a deep dive into automated bug fixing utilizing LLMs. In contrast to many deep learning-based APR methods that assume known bug locations, rely on line-level localization tools, or address bug prediction and fixing in one step, our approach uniquely employs LLMs to predict bug location at the token level and subsequently utilizes them for bug fixing. This methodological separation of bug localization and fixing using different LLMs enables effective integration of diverse contextual information and improved incorporation of inductive biases. We introduce Toggle: Token-Granulated Bug Localization and Repair, a comprehensive program repair framework that integrates a bug localization model, an adjustment unit, and a bug-fixing model. Toggle takes a buggy function as input and generates a complete corrected function. We investigate various styles of prompting to the bug fixing model to identify the most effective prompts that better utilize the inductive bias and significantly outperform others. Toggle achieves the new

Here we present DIVE: Data-driven Inference of Vertexwise Evolution. DIVE is an image-based disease progression model with single-vertex resolution, designed to reconstruct long-term patterns of brain pathology from short-term longitudinal data sets. DIVE clusters vertex-wise biomarker measurements on the cortical surface that have similar temporal dynamics across a patient population, and concurrently estimates an average trajectory of vertex measurements in each cluster. DIVE uniquely outputs a parcellation of the cortex into areas with common progression patterns, leading to a new signature for individual diseases. DIVE further estimates the disease stage and progression speed for every visit of every subject, potentially enhancing stratification for clinical trials or management. On simulated data, DIVE can recover ground truth clusters and their underlying trajectory, provided the average trajectories are sufficiently different between clusters. We demonstrate DIVE on data from two cohorts: the Alzheimer's Disease Neuroimaging Initiative (ADNI) and the Dementia Research Centre (DRC), UK, containing patients with Posterior Cortical Atrophy (PCA) as well as typical Alzheimer's d

Researchers have investigated the bug bounty ecosystem from the lens of platforms, programs, and bug hunters. Understanding the perspectives of bug bounty report reviewers, especially those who historically lack a security background and little to no funding for bug hunters, is currently understudied. In this paper, we primarily investigate the perspective of open-source software (OSS) maintainers who have used \texttt{huntr}, a bug bounty platform that pays bounties to bug hunters who find security bugs in GitHub projects and have had valid vulnerabilities patched as a result. We address this area by conducting three studies: identifying characteristics through a listing survey ($n_1=51$), their ranked importance with Likert-scale survey data ($n_2=90$), and conducting semi-structured interviews to dive deeper into real-world experiences ($n_3=17$). As a result, we categorize 40 identified characteristics into benefits, challenges, helpful features, and wanted features. We find that private disclosure and project visibility are the most important benefits, while hunters focused on money or CVEs and pressure to review are the most challenging to overcome. Surprisingly, lack of comm

This paper addresses the challenge of out-of-distribution (OOD) generalization in graph machine learning, a field rapidly advancing yet grappling with the discrepancy between source and target data distributions. Traditional graph learning algorithms, based on the assumption of uniform distribution between training and test data, falter in real-world scenarios where this assumption fails, resulting in suboptimal performance. A principal factor contributing to this suboptimal performance is the inherent simplicity bias of neural networks trained through Stochastic Gradient Descent (SGD), which prefer simpler features over more complex yet equally or more predictive ones. This bias leads to a reliance on spurious correlations, adversely affecting OOD performance in various tasks such as image recognition, natural language understanding, and graph classification. Current methodologies, including subgraph-mixup and information bottleneck approaches, have achieved partial success but struggle to overcome simplicity bias, often reinforcing spurious correlations. To tackle this, we propose DIVE, training a collection of models to focus on all label-predictive subgraphs by encouraging the

We present the first combined analysis of diving behaviour and dispersal patterns in gravid leatherback turtles during 3 consecutive nesting seasons in French Guiana. In total 23 turtles were fitted with Argos satellite transmitters and 16 individuals (including 6 concurrently satellite-tracked) were equipped with an electronic time-depth recorder for single inter-nesting intervals, i.e. between two consecutive ovi-positions. The leatherbacks dispersed over the continental shelf, ranging from the coastal zone to the shelf break and moved over 546.2 $\pm$ 154.1 km (mean $\pm$ SD) in waters of French Guiana and neighbouring Surinam. They mostly performed shallow (9.4 $\pm$ 9.2 m) and short (4.4 $\pm$ 3.4 min) dives with a slight diurnal pattern. They dived deeper as they moved away from the coast suggesting that they were predominantly following the seabed. Inter-nesting intervals could be divided into two phases: during the first 75% of the time turtles spent at sea, they dived on average 47 min h-1 before showing a lower and more variable diving effort as they came back to the shore. The extended movements of leatherbacks and the fine analysis of dive shapes suggest that in French

Objective: To study to what extent or durations of ventilation effect in a single no-decompression dive of 12 meters to a diver. Methods: There are 29 healthy volunteers divers assigned into SCUBA diving of 12m-depth underwater (the Experimental Group, EG)and chamber dive under 2.2 ATA for 20min (the Control Group, CG) matched with the factors of the age,gender,BMI and Forced Vital Capacity (FVC).Ventilation functions were measured by spirometer before diving and in 1h and 24h of post-hyperbaric exposure. Used independent samples T tests to compare the differences between the EG and CG.Analyzed of variance through repeated measurement data of different time point before or after high pressure exposure by SPSS 20.0. Results: The Inspiratory Reserve Volume(IRV) rises while the Expiratory Reserve Volume(ERV) falls significantly in 1h after high pressure release(p<0.05).So as with the Inspiratory Capacity (IC) and the Vital Capacity (VC) increased accordingly. The Ratio of FEV1.0 to VC (FEV1.0%t) is higher in CG than EG (t=-2.189,p=0.033) due to the change of VC. But the effects did not last for 24 h after high pressure relief. Conclusions: Ventilation is restricted during the 20min