ICLR 2026 论文列表

Large language models (LLMs) have demonstrated impressive capabilities in various language-based reasoning tasks (e.g., math reasoning). Among all LLM reasoning frameworks, _multi-agent debate_ (MAD), which employs multiple LLM agents and performs reasoning in the way of multi-round debate, has emerged as a powerful reasoning paradigm since it allows agents to access previous memories to refine their reasoning iteratively in each debate round and facilitates LLMs in alleviating the potential intrinsic self-preference bias. Although MAD improves the reasoning capabilities of LLMs significantly, in this paper, however, we theoretically demonstrate that the performance of MAD is closely related to the quality of memories. This indicates that MAD is still vulnerable to wrong reasoning memories, which poses a threat to the robustness of MAD. To address this problem, we introduce a simple yet effective multi-agent debate framework, _multi-agent debate with memory masking_ (MAD-M$^2$), to enhance the robustness of MAD by allowing LLM agents to select memories in the previous debate round before they perform reasoning in the current debate round. In this way, MAD-M$^2$ can polish the contextual information at the beginning of each debate round by preserving as many informative and meaningful memories as possible while dropping the noisy memories and, in turn, achieve better reasoning results. Extensive empirical results on several mainstream mathematical and logical reasoning benchmarks demonstrate that MAD-M$^2$ is able to achieve better results than the typical MAD.

Multi-Agent Discussion (MAD) has garnered increasing attention very recently, where multiple LLM instances collaboratively solve problems via structured discussion. However, we find that current MAD methods easily suffer from discussion inconsistency—LLMs fail to reach a coherent solution—due to the misalignment between their individual contexts. In this paper, we introduce a multi-LLM context learning method (M2CL) that learns a context generator for each agent, capable of dynamically generating context instructions per discussion round via automatic information organization and refinement. Specifically, inspired by our theoretical insights on the context instruction, M2CL trains the generators to control context coherence and output discrepancies via a carefully crafted self-adaptive mechanism. It enables LLMs to avoid premature convergence on “majority noise” and progressively reach the correct consensus. We evaluate M2CL on challenging tasks, including academic reasoning, embodied tasks, and mobile control. The results show that the performance of M2CL significantly surpasses existing methods by 20\%--50\%, while enjoying favorable transferability and computational efficiency.\footnote{Code is available at \url{https://github.com/HansenHua/M2CL-ICLR26}.}

Model merging has recently emerged as a lightweight alternative to ensembling, combining multiple fine-tuned models into a single set of parameters with no additional training overhead. Yet, existing merging methods fall short of matching the full accuracy of separately fine-tuned endpoints. We present MASS (MoErging through Adaptive Subspace Selection), a new approach that closes this gap by unifying multiple fine-tuned models while retaining near state-of-the-art performance across tasks. Building on the low-rank decomposition of per-task updates, MASS stores only the most salient singular components for each task and merges them into a shared model. At inference time, a non-parametric, data-free router identifies which subspace (or combination thereof) best explains an input's intermediate features and activates the corresponding task-specific block. This procedure is fully training-free and introduces only a two-pass inference overhead plus a ~2 storage factor compared to a single pretrained model, irrespective of the number of tasks. We evaluate MASS on CLIP-based image classification using ViT-B-16, ViT-B-32 and ViT-L-14 for benchmarks of 8, 14 and 20 tasks respectively, establishing a new state-of-the-art. Most notably, MASS recovers up to ~98% of the average accuracy of individual fine-tuned models, making it a practical alternative to ensembling at a fraction of the storage cost.

When a new release of a foundation model is published, practitioners typically need to repeat fine-tuning, even if the same task was already tackled in the previous version. A promising alternative is to reuse the parameter changes (i.e., task vectors) that capture how a model adapts to a specific task. However, these vectors often fail to transfer across different pre-trained models because their parameter spaces are misaligned. In this work, we show that successful transfer depends strongly on the gradient-sign structure of the new model. Based on this insight, we propose GradFix, which approximates the ideal sign structure and leverages it to transfer knowledge using only a handful of labeled samples. Notably, this requires no additional fine-tuning: we only compute a few target-model gradients without parameter updates and mask the source task vector accordingly. This yields an update that is locally aligned with the target loss landscape, effectively rebasing the task vector onto the new pre-training. We provide a theoretical guarantee that our method ensures first-order descent. Empirically, we demonstrate significant performance gains on vision and language benchmarks, consistently outperforming naive task vector addition and few-shot fine-tuning. We further show that transporting task vectors improves multi-task and multi-source model merging. Code is available at https://github.com/fillo-rinaldi/GradFix.

Model merging aims to integrate multiple independently fine-tuned expert models into a single model while preserving the knowledge of all experts. However, existing approaches mainly address parameter conflicts at the merging stage and overlook the role of the fine-tuning process, which often leads to significant post-merge performance degradation. To address this limitation, we propose a novel merging-aware optimizer (abbreviated as MergOPT) that injects principled merge-induced parameter shifts into the weight update steps so that the fine-tuned model exhibits a more stable loss landscape under subsequent merging operations. Specifically, we first formulate model merging as a distributionally robust optimization problem in the weight space: the parameters of other experts to be merged are viewed as adversarial merge-offsets, and fine-tuning adapts to the worst-case merging scenario. Building on this formulation, we analyze the distribution of parameter updates and the effects of merging hyperparameters, from which we derive a merging-guided feasible region for weight shifts. Finally, extensive experiments across four large language models (LLMs) and one vision model show that our approach consistently outperforms standard fine-tuning, yielding an average relative gain of 3.5\% and a maximum gain of 9.5\% across four merging strategies when merging seven experts.

Although agents based on multimodal large language models (MLLMs) demonstrate proficiency in general short-term graphical user interface (GUI) tasks, their robustness remains a significant challenge for handling complex long-horizon tasks in dynamic environments . In response, the LongHorizonUI framework is proposed to improve the sustained reliability of agents in long-horizon GUI tasks. To overcome core limitations, we establish a comprehensive long-horizon benchmark, LongGUIBench, covering multiple categories of games and complex general applications, with long-horizon tasks defined as requiring more than 15 steps for rigorous evaluation of long-horizon reasoning capabilities. Based on this, a Multimodal Enhanced Perceiver is designed to incorporate element detection and text recognition models, assigning unique indices to interface elements, thereby reinforcing state representation. Furthermore, a Deep Reflection Decider engine is introduced, incorporating a structured multi-level feedback validation mechanism to enable progressive reasoning and ensure accurate action execution with predictable trajectories. Finally, we introduce a Compensatory Action Executor that combines multiple degradation compensation operations with a process rollback strategy based on execution progress monitoring to ensure operational effectiveness in long-horizon task logic. Experimental results demonstrate that LongHorizonUI achieves substantial long-horizon modeling improvements on LongGUIBench while retaining competitive performance on diverse public benchmarks. The code and models will be publicly available.

Learning the dynamics of a process given sampled observations at several time points is an important but difficult task in many scientific applications. When no ground-truth trajectories are available, but one has only snapshots of data taken at discrete time steps, the problem of modelling the dynamics, and thus inferring the underlying trajectories, can be solved by multi-marginal generalisations of flow matching algorithms. This paper proposes a novel flow matching method that overcomes the limitations of existing multi-marginal trajectory inference algorithms. Our proposed method, ALI-CFM, uses a GAN-inspired adversarial loss to fit neurally parametrised interpolant curves between source and target points such that the marginal distributions at intermediate time points are close to the observed distributions. The resulting interpolants are smooth trajectories that, as we show, are unique under mild assumptions. These interpolants are subsequently marginalised by a flow matching algorithm, yielding a trained vector field for the underlying dynamics. We showcase the versatility and scalability of our method by outperforming the existing baselines on spatial transcriptomics and cell tracking datasets, while performing on par with them on single-cell trajectory prediction.

Evaluating open-ended outputs of Multimodal Large Language Models has become a bottleneck as model capabilities, task diversity, and modality rapidly expand. Existing ``MLLM-as-a-Judge'' evaluators, though promising, remain constrained to specific tasks and aspects (i.e., specific evaluation criteria such as fluency for text and image quality for images). In this paper, we argue that, on one hand, based on the interconnected nature of criteria, learning specific aspects can generalize to unseen aspects; on the other hand, jointly learning to assess multiple visual criteria and tasks may foster a synergistic effect. To this end, we propose UFEval, the first unified fine-grained evaluator with task and aspect generalization for four evaluation tasks --- Natural Language Generation, Image Understanding, Image Generation, and Interleaved Text-and-Image Generation. However, training such a unified evaluator is hindered by the lack of a large-scale, multi-modal, and aspect-level resource. To address this gap, we introduce FRABench, a comprehensive fine-grained evaluation dataset. Specifically, (1) We first construct a hierarchical aspect taxonomy encompassing 112 distinct aspects across the aforementioned four tasks. (2) Based on this taxonomy, we create FRABench, comprising 60.4k pairwise samples with 325k evaluation labels obtained from a combination of human and GPT-4o annotations. (3) Finally, leveraging FRABench, we develop UFEval, a unified fine-grained evaluator. Experiments show that learning on specific aspects enables UFEval to generalize to unseen aspects, and joint learning to assess diverse visual tasks and aspects can lead to substantial mutual benefits.

Boreal forests store 30-40\% of terrestrial carbon, much in climate-vulnerable permafrost soils, making their management critical for climate mitigation. However, optimizing forest management for both carbon sequestration and permafrost preservation presents complex trade-offs that current tools cannot adequately address. We introduce BoreaRL, the first multi-objective reinforcement learning environment for climate-adaptive boreal forest management, featuring a physically-grounded simulator of coupled energy, carbon, and water fluxes. BoreaRL supports two training paradigms: site-specific mode for controlled studies and generalist mode for learning robust policies under environmental stochasticity. Through evaluation of multi-objective RL algorithms, we reveal a fundamental asymmetry in learning difficulty: carbon objectives are significantly easier to optimize than thaw (permafrost preservation) objectives, with thaw-focused policies showing minimal learning progress across both paradigms. In generalist settings, standard gradient-descent based preference-conditioned approaches fail, while a naive site selection approach achieves superior performance by strategically selecting training episodes. Analysis of learned strategies reveals distinct management philosophies, where carbon-focused policies favor aggressive high-density coniferous stands, while effective multi-objective policies balance species composition and density to protect permafrost while maintaining carbon gains. Our results demonstrate that robust climate-adaptive forest management remains challenging for current MORL methods, establishing BoreaRL as a valuable benchmark for developing more effective approaches. We open-source BoreaRL to accelerate research in multi-objective RL for climate applications.

Learning-based methods for routing have gained significant attention in recent years, both in single-objective and multi-objective contexts. Yet, existing methods are unsuitable for routing on multigraphs, which feature multiple edges with distinct attributes between node pairs, despite their strong relevance in real-world scenarios. In this paper, we propose two graph neural network-based methods to address multi-objective routing on multigraphs. Our first approach operates directly on the multigraph by autoregressively selecting edges until a tour is completed. The second model, which is more scalable, first simplifies the multigraph via a learned pruning strategy and then performs autoregressive routing on the resulting simple graph. We evaluate both models empirically, across a wide range of problems and graph distributions, and demonstrate their competitive performance compared to strong heuristics and neural baselines.

In this paper, we study the Kronecker compressed sensing problem, which focuses on recovering sparse vectors using linear measurements obtained using the Kronecker product of two or more matrices. We first introduce the hierarchical view of the Kronecker compressed sensing, showing that the Kronecker product measurement matrix probes the sparse vector from different levels, following a block-wise and hierarchical structure. Leveraging this insight, we develop a versatile multi-stage sparse recovery algorithmic framework and tailor it to three different sparsity models: standard, hierarchical, and Kronecker-supported. We further analyze the restricted isometry property of Kronecker product matrices under different sparsity models, and provide theoretical recovery guarantees for our multi-stage algorithm. Simulations demonstrate that our method achieves comparable recovery performance to other state-of-the-art techniques while substantially reducing run time owing to the hierarchical, multi-stage recovery process.

With an ever-growing zoo of LLMs and benchmarks, the need to orchestrate multiple models for improved task performance has never been more pressing. While frameworks like Mixture-of-Agents (MoA) attempt to coordinate LLMs, they often fall short in terms of (1) selecting relevant agents, (2) facilitating effective intra-agent communication, and (3) integrating responses efficiently. In this work, we propose Graph-of-Agents (GoA), a new graph-based framework for modeling multi-agent LLM communication. Our approach begins with node sampling, selecting only the most relevant agents by leveraging model cards that summarize each model’s domain, task specialization, and other characteristics. Next, we construct edges between the selected agents by evaluating their responses against one another to determine relevance ordering. Directed message passing is then performed from highly relevant agents to less relevant ones to enhance their responses, followed by reverse message passing to refine the original responses of the more relevant agents. Finally, the updated responses are aggregated via graph-based pooling (e.g., max or mean pooling) to produce a single, unified answer. We evaluate GoA on diverse multi-domain benchmarks (MMLU, MMLU-Pro, GPQA) and domain-specific benchmarks (MATH, HumanEval, MedMCQA), with an agent pool of 6 LLMs spanning multiple domains. Surprisingly, GoA achieves superior performance18 using only 3 selected agents, outperforming recent multi-agent LLM baselines that utilize all 6 agents simultaneously. By adopting a graph structure, GoA offers both scalability and effectiveness through structured message passing—positioning it as a strong candidate for navigating the challenges of the ever-growing LLM zoo.

While the Transformer architecture dominates many fields, its quadratic self-attention complexity hinders its use in large-scale applications. **Linear attention** offers an efficient alternative, but its direct application often degrades performance, with existing fixes typically re-introducing computational overhead through extra modules (e.g., depthwise separable convolution and few self-attention blocks) that defeat the original purpose. In this work, we identify a key failure mode in these methods: **global context collapse**, where the model loses representational diversity. To address this, we propose **Multi-Head Linear Attention (MHLA)**, which preserves this diversity by computing attention within divided heads along the token dimension. We prove that MHLA maintains linear complexity while recovering much of the expressive power of softmax attention, and verify its effectiveness across multiple domains, achieving a **3.6%** improvement on ImageNet classification, a **6.3%** gain on NLP, a **12.6%** improvement in image generation tasks and a **41%** enhancement in video generation tasks with the same computational complexity,

Task Arithmetic yields a modular, scalable way to adapt foundation models. Combining multiple task vectors, however, can lead to cross-task interference, causing representation drift and degraded performance. Representation drift regularization provides a natural remedy to disentangle task vectors; however, existing approaches typically require external task data, conflicting with modularity and data availability constraints (e.g., privacy requirements). We propose a dataless approach by framing regularization against representation drift as a curvature matrix approximation problem. This allows us to leverage well-established techniques; in particular, we adopt Kronecker-Factored Approximate Curvature and obtain a practical regularizer that achieves state-of-the-art results in task addition and negation. Our method has constant complexity in the number of tasks and promotes robustness to task vector rescaling, eliminating the need for held-out tuning.

Reinforcement learning for safety-critical tasks requires policies that are both high-performing and safe throughout the learning process. While model-predictive shielding is a promising approach, existing methods are often computationally intractable for the high-dimensional, nonlinear systems where deep RL excels, as they typically rely on a patchwork of local models. We introduce **RAMPS**, a scalable shielding framework that overcomes this limitation by leveraging a learned, linear representation of the environment's dynamics. This model can range from a linear regression in the original state space to a more complex operator learned in a high-dimensional feature space. The key is that this linear structure enables a robust, look-ahead safety technique based on a *multi-step Control Barrier Function (CBF)*. By moving beyond myopic one-step formulations, **RAMPS** accounts for model error and control delays to provide reliable, real-time interventions. The resulting framework is minimally invasive, computationally efficient, and built upon robust control-theoretic foundations. Our experiments demonstrate that **RAMPS** significantly reduces safety violations compared to existing safe RL methods while maintaining high task performance in complex control environments.

Multi-task learning often suffers from gradient conflicts, leading to unfair optimization and degraded overall performance. To address this, we present SVFair, a Shapley value-based framework for fair gradient aggregation. We propose two scalable geometric conflict metrics: VolDet, a gram determinant volume metric, and VolDetPro, its sign-aware extension distinguishing antagonistic gradients. By integrating these metrics into Shapley value computation, SVFair quantifies each task’s deviation from the overall gradient and rebalances updates toward fairness. In parallel, our Shapley value computation admits controllable complexity. Extensive experiments show that SVFair achieves state-of-the-art results across diverse supervised and reinforcement learning benchmarks, and further improves existing methods when integrated as a fairness-enhancing module.

Hard constraints in reinforcement learning (RL) often degrade policy performance. Lagrangian methods offer a way to blend objectives with constraints, but require intricate reward engineering and parameter tuning. In this work, we extend recent advances that connect Hamilton-Jacobi (HJ) equations with RL to propose two novel value functions for dual-objective satisfaction. Namely, we address: 1) the Reach-Always-Avoid (RAA) problem – of achieving distinct reward and penalty thresholds – and 2) the Reach-Reach (RR) problem – of achieving thresholds of two distinct rewards. In contrast with temporal logic approaches, which typically involve representing an automaton, we derive explicit, tractable Bellman forms in this context via decomposition. Specifically, we prove that the RAA and RR problems may be rewritten as compositions of previously studied HJ-RL problems. We leverage our analysis to propose a variation of Proximal Policy Optimization (DO-HJ-PPO), and demonstrate that it produces distinct behaviors from previous approaches, out-competing a number of baselines in success, safety and speed across a range of tasks for safe-arrival and multi-target achievement.

方向：多源 RAG + 重排序优化 解决多源 RAG 中统一聚合或静态选择单一来源的局限，提出梯度感知的重排序优化方法。