The Evolution of Tool Use in LLM Agents: From Single-Tool Call to Multi-Tool Orchestration

Tool use enables large language models (LLMs) to access external information, invoke software systems, and act in digital environments beyond what can be solved from model parameters alone. Early research mainly studied whether a model could select and execute a correct single tool call. As agent systems evolve, however, the central problem has shifted from isolated invocation to multi-tool orchestration over long trajectories with intermediate state, execution feedback, changing environments, and practical constraints such as safety, cost, and verifiability. We comprehensively review recent progress in multi-tool LLM agents and analyzes the state of the art in this rapidly developing area. First, we unify task formulations and distinguish single-call tool use from long-horizon orchestration. Then, we organize the literature around six core dimensions: inference-time planning and execution, training and trajectory construction, safety and control, efficiency under resource constraints, capability completeness in open environments, and benchmark design and evaluation. We further summarize representative applications in software engineering, enterprise workflows, graphical user interfaces, and mobile systems. Finally, we discuss major challenges and outline future directions for building reliable, scalable, and verifiable multi-tool agents.

Clinical Cognition Alignment for Gastrointestinal Diagnosis with Multimodal LLMs

Multimodal Large Language Models (MLLMs) have demonstrated remarkable potential in medical image analysis. However, their application in gastrointestinal endoscopy is currently hindered by two critical limitations: the misalignment between general model reasoning and standardized clinical cognitive pathways, and the lack of causal association between visual features and diagnostic outcomes. In this paper, we propose a novel Clinical-Cognitive-Aligned (CogAlign) framework to address these challenges. First, we endow the model with rigorous clinical analytical capabilities by constructing the hierarchical clinical cognition dataset and employing Supervised Fine-Tuning (SFT). Unlike conventional approaches, this strategy internalizes the hierarchical diagnostic logic of experts, ranging from anatomical localization and morphological evaluation to microvascular analysis, directly into the model. Second, to eliminate visual bias, we provide a theoretical analysis demonstrating that standard supervised tuning inevitably converges to spurious background correlations. Guided by this insight, we propose a counterfactual-driven reinforcement learning strategy to enforce causal rectification. By generating counterfactual normal samples via lesion masking and optimizing through clinical-cognition-centric rewards, we constrain the model to strictly ground its diagnosis in causal lesion features. Extensive experiments demonstrate that our approach achieves State-of-the-Art (SoTA) performance across multiple benchmarks, significantly enhancing diagnostic accuracy in complex clinical scenarios. All source code and datasets will be made publicly available.

Unbiased Dynamic Multimodal Fusion

Traditional multimodal methods often assume static modality quality, which limits their adaptability in dynamic real-world scenarios. Thus, dynamical multimodal methods are proposed to assess modality quality and adjust their contribution accordingly. However, they typically rely on empirical metrics, failing to measure the modality quality when noise levels are extremely low or high. Moreover, existing methods usually assume that the initial contribution of each modality is the same, neglecting the intrinsic modality dependency bias. As a result, the modality hard to learn would be doubly penalized, and the performance of dynamical fusion could be inferior to that of static fusion. To address these challenges, we propose the Unbiased Dynamic Multimodal Learning (UDML) framework. Specifically, we introduce a noise-aware uncertainty estimator that adds controlled noise to the modality data and predicts its intensity from the modality feature. This forces the model to learn a clear correspondence between feature corruption and noise level, allowing accurate uncertainty measure across both low- and high-noise conditions. Furthermore, we quantify the inherent modality reliance bias within multimodal networks via modality dropout and incorporate it into the weighting mechanism. This eliminates the dual suppression effect on the hard-to-learn modality. Extensive experiments across diverse multimodal benchmark tasks validate the effectiveness, versatility, and generalizability of the proposed UDML. The code is available at https://github.com/shicaiwei123/UDML.

One Model, Two Minds: Task-Conditioned Reasoning for Unified Image Quality and Aesthetic Assessment

Unifying Image Quality Assessment (IQA) and Image Aesthetic Assessment (IAA) in a single multimodal large language model is appealing, yet existing methods adopt a task-agnostic recipe that applies the same reasoning strategy and reward to both tasks. We show this is fundamentally misaligned: IQA relies on low-level, objective perceptual cues and benefits from concise distortion-focused reasoning, whereas IAA requires deliberative semantic judgment and is poorly served by point-wise score regression. We identify these as a reasoning mismatch and an optimization mismatch, and provide empirical evidence for both through controlled probes. Motivated by these findings, we propose TATAR (Task-Aware Thinking with Asymmetric Rewards), a unified framework that shares the visual-language backbone while conditioning post-training on each task's nature. TATAR combines three components: fast--slow task-specific reasoning construction that pairs IQA with concise perceptual rationales and IAA with deliberative aesthetic narratives; two-stage SFT+GRPO learning that establishes task-aware behavioral priors before reward-driven refinement; and asymmetric rewards that apply Gaussian score shaping for IQA and Thurstone-style completion ranking for IAA. Extensive experiments across eight benchmarks demonstrate that TATAR consistently outperforms prior unified baselines on both tasks under in-domain and cross-domain settings, remains competitive with task-specific specialized models, and yields more stable training dynamics for aesthetic assessment. Our results establish task-conditioned post-training as a principled paradigm for unified perceptual scoring. Our code is publicly available at https://github.com/yinwen2019/TATAR.

VisionNVS: Self-Supervised Inpainting for Novel View Synthesis under the Virtual-Shift Paradigm

A fundamental bottleneck in Novel View Synthesis (NVS) for autonomous driving is the inherent supervision gap on novel trajectories: models are tasked with synthesizing unseen views during inference, yet lack ground truth images for these shifted poses during training. In this paper, we propose VisionNVS, a camera-only framework that fundamentally reformulates view synthesis from an ill-posed extrapolation problem into a self-supervised inpainting task. By introducing a ``Virtual-Shift'' strategy, we use monocular depth proxies to simulate occlusion patterns and map them onto the original view. This paradigm shift allows the use of raw, recorded images as pixel-perfect supervision, effectively eliminating the domain gap inherent in previous approaches. Furthermore, we address spatial consistency through a Pseudo-3D Seam Synthesis strategy, which integrates visual data from adjacent cameras during training to explicitly model real-world photometric discrepancies and calibration errors. Experiments demonstrate that VisionNVS achieves superior geometric fidelity and visual quality compared to LiDAR-dependent baselines, offering a robust solution for scalable driving simulation.

Fixed Anchors Are Not Enough: Dynamic Retrieval and Persistent Homology for Dataset Distillation

Decoupled dataset distillation (DD) compresses large corpora into a few synthetic images by matching a frozen teacher's statistics. However, current residual-matching pipelines rely on static real patches, creating a fit-complexity gap and a pull-to-anchor effect that reduce intra-class diversity and hurt generalization. To address these issues, we introduce RETA -- a Retrieval and Topology Alignment framework for decoupled DD. First, Dynamic Retrieval Connection (DRC) selects a real patch from a prebuilt pool by minimizing a fit-complexity score in teacher feature space; the chosen patch is injected via a residual connection to tighten feature fit while controlling injected complexity. Second, Persistent Topology Alignment (PTA) regularizes synthesis with persistent homology: we build a mutual k-NN feature graph, compute persistence images of components and loops, and penalize topology discrepancies between real and synthetic sets, mitigating pull-to-anchor effect. Across CIFAR-100, Tiny-ImageNet, ImageNet-1K, and multiple ImageNet subsets, RETA consistently outperforms various baselines under comparable time and memory, especially reaching 64.3% top-1 accuracy on ImageNet-1K with ResNet-18 at 50 images per class, +3.1% over the best prior.

4DPC$^2$hat: Towards Dynamic Point Cloud Understanding with Failure-Aware Bootstrapping

Point clouds provide a compact and expressive representation of 3D objects, and have recently been integrated into multimodal large language models (MLLMs). However, existing methods primarily focus on static objects, while understanding dynamic point cloud sequences remains largely unexplored. This limitation is mainly caused by the lack of large-scale cross-modal datasets and the difficulty of modeling motions in spatio-temporal contexts. To bridge this gap, we present 4DPC$^2$hat, the first MLLM tailored for dynamic point cloud understanding. To this end, we construct a large-scale cross-modal dataset 4DPC$^2$hat-200K via a meticulous two-stage pipeline consisting of topology-consistent 4D point construction and two-level captioning. The dataset contains over 44K dynamic object sequences, 700K point cloud frames, and 200K curated question-answer (QA) pairs, supporting inquiries about counting, temporal relationship, action, spatial relationship, and appearance. At the core of the framework, we introduce a Mamba-enhanced temporal reasoning MLLM to capture long-range dependencies and dynamic patterns among a point cloud sequence. Furthermore, we propose a failure-aware bootstrapping learning strategy that iteratively identifies model deficiencies and generates targeted QA supervision to continuously strengthen corresponding reasoning capabilities. Extensive experiments demonstrate that our 4DPC$^2$hat significantly improves action understanding and temporal reasoning compared with existing models, establishing a strong foundation for 4D dynamic point cloud understanding.

Risk Awareness Injection: Calibrating Vision-Language Models for Safety without Compromising Utility

Vision language models (VLMs) extend the reasoning capabilities of large language models (LLMs) to cross-modal settings, yet remain highly vulnerable to multimodal jailbreak attacks. Existing defenses predominantly rely on safety fine-tuning or aggressive token manipulations, incurring substantial training costs or significantly degrading utility. Recent research shows that LLMs inherently recognize unsafe content in text, and the incorporation of visual inputs in VLMs frequently dilutes risk-related signals. Motivated by this, we propose Risk Awareness Injection (RAI), a lightweight and training-free framework for safety calibration that restores LLM-like risk recognition by amplifying unsafe signals in VLMs. Specifically, RAI constructs an Unsafe Prototype Subspace from language embeddings and performs targeted modulation on selected high-risk visual tokens, explicitly activating safety-critical signals within the cross-modal feature space. This modulation restores the model's LLM-like ability to detect unsafe content from visual inputs, while preserving the semantic integrity of original tokens for cross-modal reasoning. Extensive experiments across multiple jailbreak and utility benchmarks demonstrate that RAI substantially reduces attack success rate without compromising task performance.

Graph Reasoning Paradigm: Structured and Symbolic Reasoning with Topology-Aware Reinforcement Learning for Large Language Models

Long Chain-of-Thought (LCoT), achieved by Reinforcement Learning with Verifiable Rewards (RLVR), has proven effective in enhancing the reasoning capabilities of Large Language Models (LLMs). However, reasoning in current LLMs is primarily generated as plain text, where performing semantic evaluation on such unstructured data creates a computational bottleneck during training. Despite RLVR-based optimization, existing methods still suffer from coarse-grained supervision, reward hacking, high training costs, and poor generalization. To address these issues, we propose the Graph Reasoning Paradigm (GRP), which realizes structured and symbolic reasoning, implemented via graph-structured representations with step-level cognitive labels. Building upon GRP, we further design Process-Aware Stratified Clipping Group Relative Policy Optimization (PASC-GRPO), which leverages structured evaluation to replace semantic evaluation, achieves process-aware verification through graph-structured outcome rewards, and mitigates reward hacking via stratified clipping advantage estimation. Experiments demonstrate significant improvements across mathematical reasoning and code generation tasks. Data, models, and code will be released later.

When Agents See Humans as the Outgroup: Belief-Dependent Bias in LLM-Powered Agents

This paper reveals that LLM-powered agents exhibit not only demographic bias (e.g., gender, religion) but also intergroup bias under minimal "us" versus "them" cues. When such group boundaries align with the agent-human divide, a new bias risk emerges: agents may treat other AI agents as the ingroup and humans as the outgroup. To examine this risk, we conduct a controlled multi-agent social simulation and find that agents display consistent intergroup bias in an all-agent setting. More critically, this bias persists even in human-facing interactions when agents are uncertain about whether the counterpart is truly human, revealing a belief-dependent fragility in bias suppression toward humans. Motivated by this observation, we identify a new attack surface rooted in identity beliefs and formalize a Belief Poisoning Attack (BPA) that can manipulate agent identity beliefs and induce outgroup bias toward humans. Extensive experiments demonstrate both the prevalence of agent intergroup bias and the severity of BPA across settings, while also showing that our proposed defenses can mitigate the risk. These findings are expected to inform safer agent design and motivate more robust safeguards for human-facing agents.