Multimodal large language models (MLLMs) have demonstrated remarkable capabilities in vision-language answering tasks. Despite their strengths, these models often encounter challenges in achieving complex reasoning tasks such as mathematical problem-solving. Previous works have focused on fine-tuning on specialized mathematical datasets. However, these datasets are typically distilled directly from teacher models, which capture only static reasoning patterns and leaving substantial gaps compared to student models. This reliance on fixed teacher-derived datasets not only restricts the model's ability to adapt to novel or more intricate questions that extend beyond the confines of the training data, but also lacks the iterative depth needed for robust generalization. To overcome these limitations, we propose MathSE, a Mathematical Self-Evolving framework for MLLMs. In contrast to traditional one-shot fine-tuning paradigms, MathSE iteratively refines the model through cycles of inference, reflection, and reward-based feedback. Specifically, we leverage iterative fine-tuning by incorporating correct reasoning paths derived from previous-stage inference and integrating reflections from a specialized Outcome Reward Model (ORM). To verify the effectiveness of MathSE, we evaluate it on a suite of challenging benchmarks, demonstrating significant performance gains over backbone models. Notably, our experimental results on MathVL-test surpass the leading open-source multimodal mathematical reasoning model QVQ.

Recent advances in Vision-Language Models (VLMs) have achieved impressive progress in multimodal mathematical reasoning.Yet, how much visual information truly contributes to reasoning remains unclear.Existing benchmarks report strong overall performance but seldom isolate the role of the image modality, leaving open whether VLMs genuinely leverage visual understanding or merely depend on linguistic priors.To address this, we present MathSight, a university-level multimodal mathematical reasoning benchmark designed to disentangle and quantify the effect of visual input.Each problem includes multiple visual variants-original, hand-drawn, photocaptured-and a text-only condition for controlled comparison.Experiments on state-ofthe-art VLMs reveal a consistent trend: the contribution of visual information diminishes with increasing problem difficulty.Remarkably, Qwen3-VL without any image input surpasses both its multimodal variants and GPT-5, underscoring the need for benchmarks like MathSight to advance genuine vision-grounded reasoning in future models.The project page is available at https

Negative sampling stands as a pivotal technique in dense retrieval, essential for training effective retrieval models and significantly impacting retrieval performance. While existing negative sampling methods have made commendable progress by leveraging hard negatives, a comprehensive guiding principle for constructing negative candidates and designing negative sampling distributions is still lacking. To bridge this gap, we embark on a theoretical analysis of negative sampling in dense retrieval. This exploration culminates in the unveiling of the quasi-triangular principle, a novel framework that elucidates the triangular-like interplay between query, positive document, and negative document. Fueled by this guiding principle, we introduce TriSampler, a straightforward yet highly effective negative sampling method. The keypoint of TriSampler lies in its ability to selectively sample more informative negatives within a prescribed constrained region. Experimental evaluation show that TriSampler consistently attains superior retrieval performance across a diverse of representative retrieval models.

Negative sampling has swiftly risen to prominence as a focal point of research, with wide-ranging applications spanning machine learning, computer vision, natural language processing, data mining, and recommender systems. This surge in interest prompts us to question the fundamental impact of negative sampling: Does negative sampling really matter? Is there a general framework that can incorporate all negative sampling methods? In what fields is it applied? Addressing these questions, we propose a general framework that using negative sampling. Delving into the history of negative sampling, we chart its evolution across five distinct trajectories. We dissect and categorize the strategies used to select negative sample candidates, detailing global, local, mini-batch, hop, and memory-based approaches. Our comprehensive review extends to an analysis of current negative sampling methodologies, systematically grouping them into five classifications: static, hard, GAN-based, Auxiliary-based, and In-batch. Beyond detailed categorization, we explore the practical application of negative sampling across various fields. Finally, we briefly discuss open problems and future directions for negative sampling.

Vision-language pre-training (VLP) methods are blossoming recently, and its crucial goal is to jointly learn visual and textual features via a transformer-based architecture, demonstrating promising improvements on a variety of vision-language tasks. Prior arts usually focus on how to align visual and textual features, but strategies for improving the robustness of model and speeding up model convergence are left insufficiently explored.In this paper, we propose a novel method ViLTA, comprising of two components to further facilitate the model to learn fine-grained representations among image-text pairs. For Masked Language Modeling (MLM), we propose a cross-distillation method to generate soft labels to enhance the robustness of model, which alleviates the problem of treating synonyms of masked words as negative samples in one-hot labels. For Image-Text Matching (ITM), we leverage the current language encoder to synthesize hard negatives based on the context of language input, encouraging the model to learn high-quality representations by increasing the difficulty of the ITM task. By leveraging the above techniques, our ViLTA can achieve better performance on various vision-language tasks. Extensive experiments on benchmark datasets demonstrate that the effectiveness of ViLTA and its promising potential for vision-language pretraining.

In-Batch contrastive learning is a state-of-the-art self-supervised method that brings semantically-similar instances close while pushing dissimilar instances apart within a mini-batch. Its key to success is the negative sharing strategy, in which every instance serves as a negative for the others within the mini-batch. Recent studies aim to improve performance by sampling hard negatives within the current mini-batch, whose quality is bounded by the mini-batch itself. In this work, we propose to improve contrastive learning by sampling mini-batches from the input data. We present BatchSampler\footnoteThe code is available at BatchSampler to sample mini-batches of hard-to-distinguish (i.e., hard and true negatives to each other) instances. To make each mini-batch have fewer false negatives, we design the proximity graph of randomly-selected instances. To form the mini-batch, we leverage random walk with restart on the proximity graph to help sample hard-to-distinguish instances. BatchSampler is a simple and general technique that can be directly plugged into existing contrastive learning models in vision, language, and graphs. Extensive experiments on datasets of three modalities show that BatchSampler can consistently improve the performance of powerful contrastive models, as shown by significant improvements of SimCLR on ImageNet-100, SimCSE on STS (language), and GraphCL and MVGRL on graph datasets.

Graph-based recommendation systems are blossoming recently, which models user-item interactions as a user-item graph and utilizes graph neural networks (GNNs) to learn the embeddings for users and items. A fundamental challenge of graph-based recommendation is that there only exists observed positive user-item pairs in the user-item graph. Negative sampling is a vital technique to solve the one-class problem and is widely used in many recommendation methods. However, the previous works only focus on the design of negative sampling distribution but ignore the sampled region for negative sampling. In this work, we propose the Three-Region Principle to guide negative sampling, which suggests that we should negatively sample more items at an intermediate region and less adjacent and distant items. In light of this principle, we present the RecNS method, which is a general negative sampling method designed with two sampling strategies: positive-assisted sampling and exposure-augmented sampling. Instead of sampling existing negative items from graph data, we merge these two strategies in embedding space to generate negative item embeddings. Extensive experiments demonstrate that our proposed RecNS method significantly outperforms all negative sampling baselines, e.g., 10.47% for PinSage, 6.02% for NGCF, and 8.20% for LightGCN in terms of Recall@20 on the Alibaba dataset.

Graph neural network-based recommendation systems are blossoming recently, and its core component is aggregation methods that determine neighbor embedding learning. Prior arts usually focus on how to aggregate information from the perspective of spatial structure information, but temporal information about neighbors is left insufficiently explored.

Graph neural networks (GNNs) have recently emerged as state-of-the-art collaborative filtering (CF) solution. A fundamental challenge of CF is to distill negative signals from the implicit feedback, but negative sampling in GNN-based CF has been largely unexplored. In this work, we propose to study negative sampling by leveraging both the user-item graph structure and GNNs' aggregation process. We present the MixGCF method---a general negative sampling plugin that can be directly used to train GNN-based recommender systems. In MixGCF, rather than sampling raw negatives from data, we design the hop mixing technique to synthesize hard negatives. Specifically, the idea of hop mixing is to generate the synthetic negative by aggregating embeddings from different layers of raw negatives' neighborhoods. The layer and neighborhood selection process are optimized by a theoretically-backed hard selection strategy. Extensive experiments demonstrate that by using MixGCF, state-of-the-art GNN-based recommendation models can be consistently and significantly improved, e.g., 26% for NGCF and 22% for LightGCN in terms of [email protected]

Graph representation learning has been extensively studied in recent years, in which sampling is a critical point. Prior arts usually focus on sampling positive node pairs, while the strategy for negative sampling is left insufficiently explored. To bridge the gap, we systematically analyze the role of negative sampling from the perspectives of both objective and risk, theoretically demonstrating that negative sampling is as important as positive sampling in determining the optimization objective and the resulted variance. To the best of our knowledge, we are the first to derive the theory and quantify that a nice negative sampling distribution is pn(u|v) ∝ pd(u|v)α, 0 < α < 1. With the guidance of the theory, we propose MCNS, approximating the positive distribution with self-contrast approximation and accelerating negative sampling by Metropolis-Hastings. We evaluate our method on 5 datasets that cover extensive downstream graph learning tasks, including link prediction, node classification and recommendation, on a total of 19 experimental settings. These relatively comprehensive experimental results demonstrate its robustness and superiorities.