ICLR 2026 Papers

Neural combinatorial optimization (NCO) has made significant advances in applying deep learning techniques to efficiently and effectively solve single-objective flexible job shop scheduling problems (FJSPs). However, the more practical multi-objective FJSPs (MOFJSPs) remain underexplored, limiting the applicability of NCO in multi-criteria decision-making scenarios. In this paper, we propose a decomposition-based NCO method to solve MOFJSPs. We present the dual conditional attention network (DCAN), a neural network architecture that takes the objective preferences along with the problem instance, aiming to learn adaptable policies over the preferences. By decomposing an MOFJSP into a set of subproblems with different preferences, the learned DCAN policies generate a set of solutions that reflect the corresponding trade-offs. We customize the Proximal Policy Optimization algorithm based on decomposition to effectively train the policy network for multiple objectives and define the state and reward based on combinations of different objectives. Extensive results showcase that our approach outperforms traditional multi-objective optimization methods and generalizes well across diverse types of problem instances.

Protein design stands as one of biology’s most important frontiers, with the potential to transform medicine, advance human health, and drive sustainability. Protein generation, a central task in protein design, has been greatly accelerated by AI-driven models—such as FoldFlow, MultiFlow, and AlphaFlow that build on residue-wise rigidity–based modeling pioneered by AlphaFold2. Residue-wise rigid-body representations reduce structural dimensionality while enforcing chemical constraints, enabling more efficient and physically consistent protein structure generation than all-atom modeling. Despite these advances, existing models often underutilize the vast structural information available in large-scale protein datasets. This highlights the importance of pretraining, which can provide richer representations and improve generalization across diverse protein design tasks. More importantly, the challenge lies in how to fully exploit abundant, low-cost unlabeled protein datasets using unsupervised pretraining. We introduce RigidSSL, a rigidity-based pretraining framework for proteins. RigidSSL canonicalizes structures into an inertial frame, employs a two-phase workflow combining large-scale perturbations and molecular dynamics views, and applies a rigid-body flow matching objective with Invariant Point Attention to capture global geometry. This enables learning stable, geometry-aware representations that improve downstream protein generation. To evaluate the effectiveness of RigidSSL, we conduct quantitative experiments on the protein generation task. Empirically, RigidSSL outperforms previous state-of-the-art geometric pretraining algorithms, leading to improvements in unconditional generation across all metrics, including designability, novelty, and diversity, for length up to 800 residues.

We introduce AnyUp, a method for feature upsampling that can be applied to any vision feature at any resolution, without encoder-specific training. Existing learning-based upsamplers for features like DINO or CLIP need to be re-trained for every feature extractor and thus do not generalize to different feature types at inference time. In this work, we propose an *inference-time* feature-agnostic upsampling architecture to alleviate this limitation and improve upsampling quality. In our experiments, AnyUp sets a new state of the art for upsampled features, generalizes to different feature types, and preserves feature semantics while being efficient and easy to apply to a wide range of downstream tasks.

Test-time scaling improves the reasoning capabilities of large language models (LLMs) by allocating extra compute to generate longer Chains-of-Thoughts (CoTs). This enables models to tackle more complex problem by breaking them down into additional steps, backtracking, and correcting mistakes. Despite its strong performance--demonstrated by OpenAI's o1 and DeepSeek R1, the conditions in the training data under which long CoTs emerge, and when such long CoTs improve the performance, remain unclear. In this paper, we study the performance of test-time scaling for transformers trained on an in-context weight prediction task for linear regression. Our analysis provides a theoretical explanation for several intriguing observations: First, at any fixed test error, increasing test-time compute allows us to reduce the number of in-context examples (context length) in training prompts. Second, if the skills required to solve a downstream task are not sufficiently present in the training data, increasing test-time compute can harm performance. Finally, we characterize task hardness via the smallest eigenvalue of its feature covariance matrix and show that training on a diverse, relevant, and hard set of tasks results in best performance for test-time scaling. We confirm our findings with experiments on large, nonlinear transformer architectures.

In-context learning (ICL) has demonstrated significant performance improvements in transformer-based large models. This study identifies two key factors influencing ICL generalisation under complex non-i.i.d. scenario: algorithmic stability and distributional discrepancy. First, we establish a stability bound for transformer-based models trained with mini-batch gradient descent, revealing how specific optimization configurations interact with the smoothness of the loss landscape to ensure the stability of non-linear Transformers. Next, we introduce a distribution-level discrepancy measure that highlights the importance of aligning the ICL prompt distribution with the training data distribution to achieve effective generalisation. Building on these insights, we derive a generalisation error bound for ICL with asymptotic convergence guarantees, which further reveals that token-wise prediction errors accumulate over time and even lead to generalisation collapse if the prediction length is not properly constrained. Finally, empirical evaluations are provided to validate our theoretical findings.

LLM-based reasoning models have enabled the development of agentic systems that act as co-scientists, assisting in multi-step scientific analysis. However, evaluating these systems is challenging, as it requires realistic, end-to-end research scenarios that integrate data analysis, interpretation, and the generation of new insights from the experimental data. To address this limitation, we introduce HeurekaBench, a framework to create benchmarks with *exploratory, open-ended research questions* for experimental datasets. Each such question is grounded in a scientific study and its corresponding code repository, and is created using a semi-automated pipeline that leverages multiple LLMs to extract insights and generate candidate workflows, which are then verified against reported findings. We instantiate the framework in single-cell biology to obtain sc-HeurekaBench benchmark and use it to compare state-of-the-art single-cell agents. We further showcase the benefits of our benchmark for quantitatively analyzing current design choices in agentic systems. We find that the addition of a *critic* module can improve ill-formed responses for open-source LLM-based agents by up to 22% and close the gap with their closed-source counterparts. Overall, HeurekaBench sets a path toward rigorous, end-to-end evaluation of scientific agents, grounding benchmark construction in real scientific workflows.

Recent advancements in Large Language Models (LLMs)-based text embedding models primarily focus on data scaling or synthesis, yet limited exploration of training techniques and data quality, thereby constraining performance. In this work, we propose KaLM-Embedding-V2, a series of versatile and compact embedding models, systematically incentivizing advanced embedding capability in LLMs by superior training techniques and high-quality data. For model architecture, we implement the models on a 0.5B compact size with simple mean-pooling to produce fixed-length embeddings and remove the causal attention mask to enable fully bidirectional representation learning. For training techniques, we propose a progressive multi-stage training pipeline: pre-training on weakly supervised large-scale datasets, fine-tuning with supervised high-quality datasets, and contrastive distillation with fine-grained soft signals, integrated with focal-style reweighting and online hard-negative mixing to emphasize difficult samples and enrich hard negatives, respectively. For training data, we curate over 20 categories for pre-training and 100 categories for fine-tuning and contrastive distillation, to improve both performance and generalization, leveraging task-specific instructions, hard-negative mining, and example-based multi-class labeling to ensure high quality. Combining these techniques, our KaLM-Embedding-V2 series achieves state-of-the-art performance on the Massive Text Embedding Benchmark, outperforming models of comparable size and rivaling models 3-26x larger, setting a new standard for versatile and compact embedding models under 1B parameters. The code, data, and models will be publicly available to facilitate academic research.

On-the-Fly Category Discovery (OCD) aims to address the limitations of transductive learning and closed-set prediction in category discovery tasks by enabling real-time classification of potential future categories using prior knowledge. Existing OCD approaches typically rely on hash-based encodings that map features into low-dimensional hash spaces and directly classify test samples using these encodings. Despite efforts to mitigate the sensitivity of hash functions during testing, these methods still suffer from severe overestimation of the number of categories. In this work, we thoroughly analyze the practical limitations of current OCD methods and formally identify a performance lower bound for the task. Based on this insight, we reformulate OCD into two sub-tasks: Open-Set Recognition and an Fully Novel OCD setting. For all samples, we employ a soft class thresholding strategy to directly detect known classes, which significantly enhances the deployment feasibility of OCD to downstream tasks. For outlier samples, we propose Adaptive Gaussian Expansion (AGE), a dynamic category discovery method that models the Probability Density Functions (PDF) of different classes to uncover potential novel categories in real time. Extensive experiments across multiple datasets demonstrate that our method achieves state-of-the-art performance.

Despite the progress in the development of generative models, their usefulness in creating synthetic data that improve prediction performance of classifiers has been put into question. Besides heuristic principles such as ''synthetic data should be close to the real data distribution'', it is actually not clear which specific properties affect the generalization error. Our paper addresses this question through the lens of high-dimensional regression. Theoretically, we show that, for linear models, the *covariance shift* between the target distribution and the distribution of the synthetic data affects the generalization error but, surprisingly, the mean shift does not. Furthermore, in some regimes, we prove that matching the covariance of the target distribution is optimal. Remarkably, the theoretical insights for linear models carry over to deep neural networks and generative models. We empirically demonstrate that the *covariance matching* procedure (matching the covariance of the synthetic data with that of the data coming from the target distribution) performs well against several recent approaches for synthetic data selection, across various training paradigms, datasets and generative models used for augmentation.

Pre-trained large language models (LLMs) show promise for robotic task planning but often struggle to guarantee correctness in long-horizon problems. Task and motion planning (TAMP) addresses this by grounding symbolic plans in low-level execution, yet it relies heavily on manually engineered planning domains. To improve long-horizon planning reliability and reduce human intervention, we present Planning Domain Derivation with LLMs (PDDLLM), a framework that automatically induces symbolic predicates and actions directly from demonstration trajectories by combining LLM reasoning with physical simulation roll-outs. Unlike prior domain-inference methods that rely on partially predefined or language descriptions of planning domains, PDDLLM constructs domains with minimal manual domain initialization and automatically integrates them with motion planners to produce executable plans, enhancing long-horizon planning automation. Across 1,200 tasks in nine environments, PDDLLM outperforms six LLM-based planning baselines, achieving at least 20% higher success rates, reduced token costs, and successful deployment on multiple physical robot platforms.

Language models have advanced sequence analysis, yet DNA foundation models often lag behind task-specific methods for unclear reasons. We present AntigenLM, a generative DNA language model pretrained on influenza genomes with intact, aligned functional units. This structure-aware pretraining enables AntigenLM to capture evolutionary constraints and generalize across tasks. Fine-tuned on time-series hemagglutinin (HA) and neuraminidase (NA) sequences, AntigenLM accurately forecasts future antigenic variants across regions and subtypes, including those unseen during training, outperforming phylogenetic and evolution-based models. It also achieves near-perfect subtype classification. Ablation studies show that disrupting genomic structure through fragmentation or shuffling severely degrades performance, revealing the importance of preserving functional-unit integrity in DNA language modeling. AntigenLM thus provides both a powerful framework for antigen evolution prediction and a general principle for building biologically grounded DNA foundation models.

Learning-based 3D reconstruction models, represented by Visual Geometry Grounded Transformers (VGGTs), have achieved remarkable progress with large-scale transformers. Their prohibitive computational and memory costs severely hinder real-world deployment. Post-Training Quantization (PTQ) has emerged as a common practice to compress and accelerate models. However, we empirically observe that PTQ faces unique obstacles when compressing billion-scale VGGTs: the data-independent special tokens induce heavy-tailed activation distributions, while the multi-view nature of 3D data makes calibration sample selection highly unstable. This paper proposes the first **Quant**ization framework for **VGGT**s, namely **QuantVGGT**. This mainly relies on two technical contributions: First, we introduce *Dual-Smoothed Fine-Grained Quantization*, which integrates pre-global Hadamard rotation and post-local channel smoothing to robustly mitigate heavy-tailed distributions and inter-channel variance. Second, we design *Noise-Filtered Diverse Sampling*, which filters outliers via deep-layer statistics and constructs frame-aware diverse calibration clusters to ensure stable quantization ranges. Comprehensive experiments demonstrate that QuantVGGT achieves the state-of-the-art results across different benchmarks and bit-width, surpassing the previous state-of-the-art generic quantization method with a great margin. We highlight that our 4-bit QuantVGGT can deliver a **3.7$\times$** memory reduction and **2.5$\times$** acceleration in real-hardware inference, while preserving over **98\%** reconstruction accuracy of the full-precision counterparts. This demonstrates the vast advantages and practicality of QuantVGGT in resource-constrained scenarios.

Apparel is a fundamental component of human appearance, making garment digitalization critical for digital human creation. However, sewing pattern creation traditionally relies on the intuition and extensive experience of skilled artisans. This manual bottleneck significantly hinders the scalability of digital garment creation. Existing generative approaches either operate as data replicators without intrinsic understanding of garment construction principles (e.g., diffusion models), or struggle with low-level regression of raw floating-point coordinates (e.g., Vision-Language Models). We present GarmentGPT, the first framework to operationalize latent space generation for sewing patterns. Our approach introduces a novel pipeline where a RVQ-VAE tokenizes continuous pattern boundary curves into discrete codebook indices. A fine-tuned Vision-Language Model then autoregressively predicts these discrete token sequences instead of regressing coordinates, enabling high-level compositional reasoning. This paradigm shift aligns generation with the knowledge-driven, symbolic reasoning capabilities of large language models. To address the data bottleneck for real-world applications, we develop a Data Curation Pipeline that synthesizes over one million photorealistic images paired with GarmentCode, and establish the Real-Garments Benchmark for comprehensive evaluation. Experiments demonstrate that GarmentGPT significantly outperforms existing methods on structured datasets (95.62\% Panel Accuracy, 81.84\% Stitch Accuracy), validating our discrete compositional paradigm's advantages. Code is available at \url{https://github.com/ChimerAI-MMLab/Garment-GPT}.

Modern neural networks exhibit a striking property: basins of attraction in the loss landscape are often connected by low-loss paths, yet optimization dynamics generally remain confined to a single convex basin and rarely explore intermediate points. We resolve this paradox by identifying entropic barriers arising from the interplay between curvature variations along these paths and noise in optimization dynamics. Empirically, we find that curvature systematically rises away from minima, producing effective forces that bias noisy dynamics back toward the endpoints — even when the loss remains nearly flat. These barriers persist longer than energetic barriers, shaping the late-time localization of solutions in parameter space. Our results highlight the role of curvature-induced entropic forces in governing both connectivity and confinement in deep learning landscapes.

Forecasting how 3D medical scans evolve along time is important for disease progression, treatment planning, and developmental assessment. Yet existing models either rely on a single prior scan, fixed grid times, or target global labels, which limits voxel-level forecasting under irregular sampling. We present CRONOS, a unified framework for many-to-one prediction from multiple past scans that supports both discrete (grid-based) and continuous (real-valued) timestamps in one model, to the best of our knowledge the first to achieve continuous sequence-to-image forecasting for 3D medical data. CRONOS learns a spatio-temporal velocity field that transports context volumes toward a target volume at an arbitrary time, while operating directly in 3D voxel space. Across three public datasets spanning Cine-MRI, perfusion CT, and longitudinal MRI, CRONOS outperforms other baselines, while remaining computationally competitive. We will release code and evaluation protocols to enable reproducible, multi-dataset benchmarking of multi-context, continuous-time forecasting.

The Multi-Head Attention mechanism is central to LLM operation, and multiple works target its compute and memory efficiency during training. While most works focus on approximating the scaled dot product, the memory consumption of the linear projections that compute the $Q$, $K$, and $V$ tensors from the input $x$ is often overlooked. To address this, we propose Point-Approximate Matrix Multiplication (PAMM), a novel tensor compression technique that reduces memory consumption of the $Q,K,V$ projections in attention layers by a factor of up to $\times 512$, effectively erasing their memory footprint, while achieving similar or better final perplexity. PAMM is fully composable with efficient attention techniques such as FlashAttention, making it a practical and complementary method for memory-efficient LLM training.

Diffusion Large Language Models (DLLMs) have emerged as a compelling alternative to Autoregressive models, designed for fast parallel generation. However, existing DLLMs are plagued by a severe quality-speed trade-off, where faster parallel decoding leads to significant performance degradation. We attribute this to the irreversibility of standard decoding in DLLMs, which is easily polarized into the wrong decoding direction along with early error context accumulation. To resolve this, we introduce Wide-In, Narrow-Out (WINO), a training-free decoding algorithm that enables revokable decoding in DLLMs. WINO employs a parallel draft-and-verify mechanism, aggressively drafting multiple tokens while simultaneously using the model's bidirectional context to verify and re-mask suspicious ones for refinement. Verified in open-source DLLMs like LLaDA and MMaDA, WINO is shown to decisively improve the quality-speed trade-off. For instance, on the GSM8K math benchmark, it accelerates inference by 6 while improving accuracy by 2.58%; on Flickr30K captioning, it achieves a 10 speedup with higher performance. More comprehensive experiments are conducted to demonstrate the superiority and provide an in-depth understanding of WINO.

Attribution methods seek to explain language model predictions by quantifying the contribution of input tokens to generated outputs. However, most existing techniques are designed for encoder-based architectures and rely on linear approximations that fail to capture the causal and semantic complexities of autoregressive generation in decoder-only models. To address these limitations, we propose **Hessian-Enhanced Token Attribution (HETA)**, a novel attribution framework tailored for decoder-only language models. HETA combines three complementary components: a semantic transition vector that captures token-to-token influence across layers, Hessian-based sensitivity scores that model second-order effects, and KL divergence to measure information loss when tokens are masked. This unified design produces context-aware, causally faithful, and semantically grounded attributions. Additionally, we introduce a **curated benchmark dataset** for systematically evaluating attribution quality in generative settings. Empirical evaluations across multiple models and datasets demonstrate that HETA consistently outperforms existing methods in attribution faithfulness and alignment with human annotations, establishing a new standard for interpretability in autoregressive language models.

Pre-trained models exhibit strong generalization to various downstream tasks. However, given the numerous models available in the model hub, identifying the most suitable one by individually fine-tuning is time-consuming. In this paper, we propose \textbf{SwiftTS}, a swift selection framework for time series pre-trained models. To avoid expensive forward propagation through all candidates, SwiftTS adopts a learning-guided approach that leverages historical dataset-model performance pairs across diverse horizons to predict model performance on unseen datasets. It employs a lightweight dual-encoder architecture that embeds time series and candidate models with rich characteristics, computing patchwise compatibility scores between data and model embeddings for efficient selection. To further enhance the generalization across datasets and horizons, we introduce a horizon-adaptive expert composition module that dynamically adjusts expert weights, and the transferable cross-task learning with cross-dataset and cross-horizon task sampling to enhance out-of-distribution (OOD) robustness. Extensive experiments on 14 downstream datasets and 8 pre-trained models demonstrate that SwiftTS achieves state-of-the-art performance in time series pre-trained model selection. The code and datasets are available at \href{}{https://anonymous.4open.science/r/SwiftTS-395C}.