Large language models (LLMs) incorporated with Retrieval-Augmented Generation (RAG) have demonstrated powerful capabilities in generating counterspeech against misinformation. However, current studies rely on limited evidence and offer less control over final outputs. To address these challenges, we propose a Multi-agent Retrieval-Augmented Framework to generate counterspeech against health misinformation, incorporating multiple LLMs to optimize knowledge retrieval, evidence enhancement, and response refinement. Our approach integrates both static and dynamic evidence, ensuring that the generated counterspeech is relevant, well-grounded, and up-to-date. Our method outperforms baseline approaches in politeness, relevance, informativeness, and factual accuracy, demonstrating its effectiveness in generating high-quality counterspeech. To further validate our approach, we conduct ablation studies to verify the necessity of each component in our framework. Furthermore, human evaluations reveal that refinement significantly enhances counterspeech quality and obtains human preference.
Recent benchmarks for Large Language Model (LLM) agents primarily focus on evaluating reasoning, planning, and execution capabilities, while another critical component-memory, encompassing how agents memorize, update, and retrieve long-term information-is under-evaluated due to the lack of benchmarks. We term agents with memory mechanisms as memory agents. In this paper, we identify four core competencies essential for memory agents: accurate retrieval, test-time learning, long-range understanding, and conflict resolution. Existing datasets either rely on limited context lengths or are tailored for static, long-context settings like book-based QA, which do not reflect the interactive, multi-turn nature of memory agents that incrementally accumulate information. Furthermore, no existing benchmarks cover all four competencies. Therefore, we introduce MemoryAgentBench, a new benchmark specifically designed for memory agents. Our benchmark combines reformulated existing datasets with newly constructed ones, covering the above four memory competencies, providing a systematic and challenging testbed for assessing memory quality. We evaluate a diverse set of memory agents, ranging from simple context-based and retrieval-augmented generation (RAG) systems to advanced agents with external memory modules and tool integration. Empirical results reveal that current methods fall short of mastering all four competencies, underscoring the need for further research into comprehensive memory mechanisms for LLM agents.
While large language models have demonstrated impressive capabilities in web navigation tasks, the extensive context of web pages, often represented as DOM or Accessibility Tree (AxTree) structures, frequently exceeds model context limits. Current approaches like bottom-up truncation or embedding-based retrieval lose critical information about page state and action history. This is particularly problematic for adaptive planning in web agents, where understanding the current state is essential for determining future actions. We hypothesize that embedding models lack sufficient capacity to capture plan-relevant information, especially when retrieving content that supports future action prediction. This raises a fundamental question: how can retrieval methods be optimized for adaptive planning in web navigation tasks? In response, we introduce \textit{LineRetriever}, a novel approach that leverages a language model to identify and retrieve observation lines most relevant to future navigation steps. Unlike traditional retrieval methods that focus solely on semantic similarity, \textit{LineRetriever} explicitly considers the planning horizon, prioritizing elements that contribute to action prediction. Our experiments demonstrate that \textit{LineRetriever} can reduce the size of the observation at each step for the web agent while maintaining consistent performance within the context limitations.
Multimodal document retrieval systems enable information access across text, images, and layouts, benefiting various domains like document-based question answering, report analysis, and interactive content summarization. Rerankers improve retrieval precision by reordering retrieved candidates. However, current multimodal reranking methods remain underexplored, with significant room for improvement in both training strategies and overall effectiveness. Moreover, the lack of explicit reasoning makes it difficult to analyze and optimize these methods further. In this paper, We propose MM-R5, a MultiModal Reasoning-Enhanced ReRanker via Reinforcement Learning for Document Retrieval, aiming to provide a more effective and reliable solution for multimodal reranking tasks. MM-R5 is trained in two stages: supervised fine-tuning (SFT) and reinforcement learning (RL). In the SFT stage, we focus on improving instruction-following and guiding the model to generate complete and high-quality reasoning chains. To support this, we introduce a novel data construction strategy that produces rich, high-quality reasoning data. In the RL stage, we design a task-specific reward framework, including a reranking reward tailored for multimodal candidates and a composite template-based reward to further refine reasoning quality. We conduct extensive experiments on MMDocIR, a challenging public benchmark spanning multiple domains. MM-R5 achieves state-of-the-art performance on most metrics and delivers comparable results to much larger models on the remaining ones. Moreover, compared to the best retrieval-only method, MM-R5 improves recall@1 by over 4%. These results validate the effectiveness of our reasoning-enhanced training pipeline. Our code is available at https://github.com/i2vec/MM-R5 .
As a challenging vision-language (VL) task, Composed Image Retrieval (CIR) aims to retrieve target images using multimodal (image+text) queries. Although many existing CIR methods have attained promising performance, their reliance on costly, manually labeled triplets hinders scalability and zero-shot capability. To address this issue, we propose a scalable pipeline for automatic triplet generation, along with a fully synthetic dataset named Composed Image Retrieval on High-quality Synthetic Triplets (CIRHS). Our pipeline leverages a large language model (LLM) to generate diverse prompts, controlling a text-to-image generative model to produce image pairs with identical elements in each pair, which are then filtered and reorganized to form the CIRHS dataset. In addition, we introduce Hybrid Contextual Alignment (CoAlign), a novel CIR framework, which can accomplish global alignment and local reasoning within a broader context, enabling the model to learn more robust and informative representations. By utilizing the synthetic CIRHS dataset, CoAlign achieves outstanding zero-shot performance on three commonly used benchmarks, demonstrating for the first time the feasibility of training CIR models on a fully synthetic dataset. Furthermore, under supervised training, our method outperforms all the state-of-the-art supervised CIR approaches, validating the effectiveness of our proposed retrieval framework. The code and the CIRHS dataset will be released soon.
Motivated by the growing demand for retrieval systems that operate across modalities, we introduce llama-nemoretriever-colembed, a unified text-image retrieval model that delivers state-of-the-art performance across multiple benchmarks. We release two model variants, 1B and 3B. The 3B model achieves state of the art performance, scoring NDCG@5 91.0 on ViDoRe V1 and 63.5 on ViDoRe V2, placing first on both leaderboards as of June 27, 2025. Our approach leverages the NVIDIA Eagle2 Vision-Language model (VLM), modifies its architecture by replacing causal attention with bidirectional attention, and integrates a ColBERT-style late interaction mechanism to enable fine-grained multimodal retrieval in a shared embedding space. While this mechanism delivers superior retrieval accuracy, it introduces trade-offs in storage and efficiency. We provide a comprehensive analysis of these trade-offs. Additionally, we adopt a two-stage training strategy to enhance the model's retrieval capabilities.
Vector retrieval systems exhibit significant performance variance across queries due to heterogeneous embedding quality. We propose a lightweight framework for predicting retrieval performance at the query level by combining quantization robustness and neighborhood density metrics. Our approach is motivated by the observation that high-quality embeddings occupy geometrically stable regions in the embedding space and exhibit consistent neighborhood structures. We evaluate our method on 4 standard retrieval datasets, showing consistent improvements of 9.4$\pm$1.2\% in Recall@10 over competitive baselines. The framework requires minimal computational overhead (less than 5\% of retrieval time) and enables adaptive retrieval strategies. Our analysis reveals systematic patterns in embedding quality across different query types, providing insights for targeted training data augmentation.
Effective cross-modal retrieval is essential for applications like information retrieval and recommendation systems, particularly in specialized domains such as manufacturing, where product information often consists of visual samples paired with a textual description. This paper investigates the use of Vision Language Models(VLMs) for zero-shot text-to-image retrieval on fabric samples. We address the lack of publicly available datasets by introducing an automated annotation pipeline that uses Multimodal Large Language Models (MLLMs) to generate two types of textual descriptions: freeform natural language and structured attribute-based descriptions. We produce these descriptions to evaluate retrieval performance across three Vision-Language Models: CLIP, LAION-CLIP, and Meta's Perception Encoder. Our experiments demonstrate that structured, attribute-rich descriptions significantly enhance retrieval accuracy, particularly for visually complex fabric classes, with the Perception Encoder outperforming other models due to its robust feature alignment capabilities. However, zero-shot retrieval remains challenging in this fine-grained domain, underscoring the need for domain-adapted approaches. Our findings highlight the importance of combining technical textual descriptions with advanced VLMs to optimize cross-modal retrieval in industrial applications.
Sequence models like Transformers and RNNs often overallocate attention to irrelevant context, leading to noisy intermediate representations. This degrades LLM capabilities by promoting hallucinations, weakening long-range and retrieval abilities, and reducing robustness. Recent work has shown that differential design can mitigate this issue in Transformers, improving their effectiveness across various applications. In this paper, we explore whether these techniques, originally developed for Transformers, can be applied to Mamba, a recent architecture based on selective state-space layers that achieves Transformer-level performance with greater efficiency. We show that a naive adaptation of differential design to Mamba is insufficient and requires careful architectural modifications. To address this, we introduce a novel differential mechanism for Mamba, empirically validated on language modeling benchmarks, demonstrating improved retrieval capabilities and superior performance over vanilla Mamba. Finally, we conduct extensive ablation studies and empirical analyses to justify our design choices and provide evidence that our approach effectively mitigates the overallocation problem in Mamba-based models. Our code is publicly available.
Large language models (LLMs), including zero-shot and few-shot paradigms, have shown promising capabilities in clinical text generation. However, real-world applications face two key challenges: (1) patient data is highly unstructured, heterogeneous, and scattered across multiple note types and (2) clinical notes are often long and semantically dense, making naive prompting infeasible due to context length constraints and the risk of omitting clinically relevant information. We introduce CLI-RAG (Clinically Informed Retrieval-Augmented Generation), a domain-specific framework for structured and clinically grounded text generation using LLMs. It incorporates a novel hierarchical chunking strategy that respects clinical document structure and introduces a task-specific dual-stage retrieval mechanism. The global stage identifies relevant note types using evidence-based queries, while the local stage extracts high-value content within those notes creating relevance at both document and section levels. We apply the system to generate structured progress notes for individual hospital visits using 15 clinical note types from the MIMIC-III dataset. Experiments show that it preserves temporal and semantic alignment across visits, achieving an average alignment score of 87.7%, surpassing the 80.7% baseline from real clinician-authored notes. The generated outputs also demonstrate high consistency across LLMs, reinforcing deterministic behavior essential for reproducibility, reliability, and clinical trust.
Retrieval-augmented Generation (RAG) extends large language models (LLMs) with external knowledge but faces key challenges: restricted effective context length and redundancy in retrieved documents. Pure compression-based approaches reduce input size but often discard fine-grained details essential for factual accuracy. We propose SARA, a unified RAG framework that balances local precision and global knowledge coverage under tight context budgets. SARA combines natural-language text snippets with semantic compression vectors to jointly enhance context efficiency and answer correctness. It represents contexts at two complementary levels: 1) fine-grained natural-language spans that preserve critical entities and numerical values, and 2) compact, interpretable vectors that summarize high-level semantics. An iterative evidence-selection module employs the compression vectors for dynamic reranking of contexts. Across 9 datasets and 5 open-source LLMs spanning 3 model families (Mistral, Llama, and Gemma), SARA consistently improves answer relevance (+17.71), answer correctness (+13.72), and semantic similarity (+15.53), demonstrating the importance of integrating textual and compressed representations for robust, context-efficient RAG.
Text-to-image retrieval (TIR) aims to find relevant images based on a textual query, but existing approaches are primarily based on whole-image captions and lack interpretability. Meanwhile, referring expression segmentation (RES) enables precise object localization based on natural language descriptions but is computationally expensive when applied across large image collections. To bridge this gap, we introduce Mask-aware TIR (MaTIR), a new task that unifies TIR and RES, requiring both efficient image search and accurate object segmentation. To address this task, we propose a two-stage framework, comprising a first stage for segmentation-aware image retrieval and a second stage for reranking and object grounding with a multimodal large language model (MLLM). We leverage SAM 2 to generate object masks and Alpha-CLIP to extract region-level embeddings offline at first, enabling effective and scalable online retrieval. Secondly, MLLM is used to refine retrieval rankings and generate bounding boxes, which are matched to segmentation masks. We evaluate our approach on COCO and D$^3$ datasets, demonstrating significant improvements in both retrieval accuracy and segmentation quality over previous methods.
Centuries-old Islamic medical texts like Avicenna's Canon of Medicine and the Prophetic Tibb-e-Nabawi encode a wealth of preventive care, nutrition, and holistic therapies, yet remain inaccessible to many and underutilized in modern AI systems. Existing language-model benchmarks focus narrowly on factual recall or user preference, leaving a gap in validating culturally grounded medical guidance at scale. We propose a unified evaluation pipeline, Tibbe-AG, that aligns 30 carefully curated Prophetic-medicine questions with human-verified remedies and compares three LLMs (LLaMA-3, Mistral-7B, Qwen2-7B) under three configurations: direct generation, retrieval-augmented generation, and a scientific self-critique filter. Each answer is then assessed by a secondary LLM serving as an agentic judge, yielding a single 3C3H quality score. Retrieval improves factual accuracy by 13%, while the agentic prompt adds another 10% improvement through deeper mechanistic insight and safety considerations. Our results demonstrate that blending classical Islamic texts with retrieval and self-evaluation enables reliable, culturally sensitive medical question-answering.
Retrieval augmented generation (RAG) has shown great power in improving Large Language Models (LLMs). However, most existing RAG-based LLMs are dedicated to retrieving single modality information, mainly text; while for many real-world problems, such as healthcare, information relevant to queries can manifest in various modalities such as knowledge graph, text (clinical notes), and complex molecular structure. Thus, being able to retrieve relevant multi-modality domain-specific information, and reason and synthesize diverse knowledge to generate an accurate response is important. To address the gap, we present BioMol-MQA, a new question-answering (QA) dataset on polypharmacy, which is composed of two parts (i) a multimodal knowledge graph (KG) with text and molecular structure for information retrieval; and (ii) challenging questions that designed to test LLM capabilities in retrieving and reasoning over multimodal KG to answer questions. Our benchmarks indicate that existing LLMs struggle to answer these questions and do well only when given the necessary background data, signaling the necessity for strong RAG frameworks.
Many promising-looking ideas in AI research fail to deliver, but their validation takes substantial human labor and compute. Predicting an idea's chance of success is thus crucial for accelerating empirical AI research, a skill that even expert researchers can only acquire through substantial experience. We build the first benchmark for this task and compare LMs with human experts. Concretely, given two research ideas (e.g., two jailbreaking methods), we aim to predict which will perform better on a set of benchmarks. We scrape ideas and experimental results from conference papers, yielding 1,585 human-verified idea pairs published after our base model's cut-off date for testing, and 6,000 pairs for training. We then develop a system that combines a fine-tuned GPT-4.1 with a paper retrieval agent, and we recruit 25 human experts to compare with. In the NLP domain, our system beats human experts by a large margin (64.4% v.s. 48.9%). On the full test set, our system achieves 77% accuracy, while off-the-shelf frontier LMs like o3 perform no better than random guessing, even with the same retrieval augmentation. We verify that our system does not exploit superficial features like idea complexity through extensive human-written and LM-designed robustness tests. Finally, we evaluate our system on unpublished novel ideas, including ideas generated by an AI ideation agent. Our system achieves 63.6% accuracy, demonstrating its potential as a reward model for improving idea generation models. Altogether, our results outline a promising new direction for LMs to accelerate empirical AI research.
Rerankers play a critical role in multimodal Retrieval-Augmented Generation (RAG) by refining ranking of an initial set of retrieved documents. Rerankers are typically trained using hard negative mining, whose goal is to select pages for each query which rank high, but are actually irrelevant. However, this selection process is typically passive and restricted to what the retriever can find in the available corpus, leading to several inherent limitations. These include: limited diversity, negative examples which are often not hard enough, low controllability, and frequent false negatives which harm training. Our paper proposes an alternative approach: Single-Page Hard Negative Query Generation, which goes the other way around. Instead of retrieving negative pages per query, we generate hard negative queries per page. Using an automated LLM-VLM pipeline, and given a page and its positive query, we create hard negatives by rephrasing the query to be as similar as possible in form and context, yet not answerable from the page. This paradigm enables fine-grained control over the generated queries, resulting in diverse, hard, and targeted negatives. It also supports efficient false negative verification. Our experiments show that rerankers trained with data generated using our approach outperform existing models and significantly improve retrieval performance.
Retrieval-Augmented Generation (RAG) has shown promise in enhancing recommendation systems by incorporating external context into large language model prompts. However, existing RAG-based approaches often rely on static retrieval heuristics and fail to capture nuanced user preferences in dynamic recommendation scenarios. In this work, we introduce ARAG, an Agentic Retrieval-Augmented Generation framework for Personalized Recommendation, which integrates a multi-agent collaboration mechanism into the RAG pipeline. To better understand the long-term and session behavior of the user, ARAG leverages four specialized LLM-based agents: a User Understanding Agent that summarizes user preferences from long-term and session contexts, a Natural Language Inference (NLI) Agent that evaluates semantic alignment between candidate items retrieved by RAG and inferred intent, a context summary agent that summarizes the findings of NLI agent, and an Item Ranker Agent that generates a ranked list of recommendations based on contextual fit. We evaluate ARAG accross three datasets. Experimental results demonstrate that ARAG significantly outperforms standard RAG and recency-based baselines, achieving up to 42.1% improvement in NDCG@5 and 35.5% in Hit@5. We also, conduct an ablation study to analyse the effect by different components of ARAG. Our findings highlight the effectiveness of integrating agentic reasoning into retrieval-augmented recommendation and provide new directions for LLM-based personalization.
Deep Research Agents are a prominent category of LLM-based agents. By autonomously orchestrating multistep web exploration, targeted retrieval, and higher-order synthesis, they transform vast amounts of online information into analyst-grade, citation-rich reports--compressing hours of manual desk research into minutes. However, a comprehensive benchmark for systematically evaluating the capabilities of these agents remains absent. To bridge this gap, we present DeepResearch Bench, a benchmark consisting of 100 PhD-level research tasks, each meticulously crafted by domain experts across 22 distinct fields. Evaluating DRAs is inherently complex and labor-intensive. We therefore propose two novel methodologies that achieve strong alignment with human judgment. The first is a reference-based method with adaptive criteria to assess the quality of generated research reports. The other framework is introduced to evaluate DRA's information retrieval and collection capabilities by assessing its effective citation count and overall citation accuracy. We have open-sourced DeepResearch Bench and key components of these frameworks at https://github.com/Ayanami0730/deep_research_bench to accelerate the development of practical LLM-based agents.
Large Language Models (LLMs) have demonstrated remarkable capabilities across various tasks, while they remain prone to generating hallucinated or outdated responses due to their static internal knowledge. Recent advancements in Retrieval-Augmented Generation (RAG) methods have explored enhancing models' search and reasoning capabilities through reinforcement learning (RL). Although these methods demonstrate promising results, they face challenges in training stability and encounter issues such as substantial inference time and restricted capabilities due to the single-query mode. In this paper, we propose RAG-R1, a novel training framework designed to enable LLMs to adaptively leverage internal and external knowledge during the reasoning process. We further expand the generation and retrieval processes within the framework from single-query mode to multi-query parallelism, aimed at reducing inference time and enhancing the model's capabilities. Extensive experiments on seven question-answering benchmarks demonstrate that our method outperforms the strongest baseline by up to 13.2% and decreases inference time by 11.1%.
The ability to retrieve semantically and directionally similar short-range trajectories with both accuracy and efficiency is foundational for downstream applications such as motion forecasting and autonomous navigation. However, prevailing approaches often depend on computationally intensive heuristics or latent anchor representations that lack interpretability and controllability. In this work, we propose a novel framework for learning fixed-dimensional embeddings for short trajectories by leveraging a Transformer encoder trained with a contrastive triplet loss that emphasize the importance of discriminative feature spaces for trajectory data. We analyze the influence of Cosine and FFT-based similarity metrics within the contrastive learning paradigm, with a focus on capturing the nuanced directional intent that characterizes short-term maneuvers. Our empirical evaluation on the Argoverse 2 dataset demonstrates that embeddings shaped by Cosine similarity objectives yield superior clustering of trajectories by both semantic and directional attributes, outperforming FFT-based baselines in retrieval tasks. Notably, we show that compact Transformer architectures, even with low-dimensional embeddings (e.g., 16 dimensions, but qualitatively down to 4), achieve a compelling balance between retrieval performance (minADE, minFDE) and computational overhead, aligning with the growing demand for scalable and interpretable motion priors in real-time systems. The resulting embeddings provide a compact, semantically meaningful, and efficient representation of trajectory data, offering a robust alternative to heuristic similarity measures and paving the way for more transparent and controllable motion forecasting pipelines.