Learning from procedural videos remains a core challenge in self-supervised representation learning, as real-world instructional data often contains background segments, repeated actions, and steps presented out of order. Such variability violates the strong monotonicity assumptions underlying many alignment methods. Prior state-of-the-art approaches, such as OPEL, leverage Kantorovich Optimal Transport (KOT) to build frame-to-frame correspondences, but rely solely on feature similarity and fail to capture the higher-order temporal structure of a task. In this paper, we introduce REALIGN, a self-supervised framework for procedure learning based on Regularized Fused Partial Gromov-Wasserstein Optimal Transport (R-FPGWOT). In contrast to KOT, our formulation jointly models visual correspondences and temporal relations under a partial alignment scheme, enabling robust handling of irrelevant frames, repeated actions, and non-monotonic step orders common in instructional videos. To stabilize training, we integrate FPGWOT distances with inter-sequence contrastive learning, avoiding the need for multiple regularizers and preventing collapse to degenerate solutions. Across egocentric (EgoProceL) and third-person (ProceL, CrossTask) benchmarks, REALIGN achieves up to 18.9% average F1-score improvements and over 30% temporal IoU gains, while producing more interpretable transport maps that preserve key-step orderings and filter out noise.

Existing methods usually leverage a fixed strategy, such as natural language reasoning, code-augmented reasoning, tool-integrated reasoning, or ensemble-based reasoning, to guide Large Language Models (LLMs) to perform mathematical reasoning. Our analysis reveals that the single strategy cannot adapt to problem-specific requirements and thus overlooks the trade-off between effectiveness and efficiency. To address these issues, we propose Planning and Routing through Instance-Specific Modeling (PRISM), a novel framework that decouples mathematical reasoning into two stages: strategy planning and targeted execution. Specifically, we first curate a multi-strategy preference dataset, which we call MathStrat, capturing correctness, process quality, and computational efficiency for each problem--strategy pair. Then, we train a lightweight Strategy Adapter based on the dataset to obtain confidence distributions over the mentioned four reasoning strategies. At inference time, an adaptive routing policy dynamically tailors the reasoning approach based on predictor confidence. It directs the model to use single-strategy execution for high-confidence predictions, dual-strategy verification for competitive scenarios, or comprehensive multi-strategy exploration for uncertain cases. Extensive experiments across five mathematical reasoning benchmarks demonstrate that PRISM consistently outperforms individual strategies and ensemble baselines, achieving improvements ranging from 0.9% to 7.6% across different base models. The adaptive routing approach shows particularly strong benefits for mathematical reasoning tasks across diverse model architectures. Our code is released at https://github.com/reml-group/PRISM.

The development of state-of-the-art large language models is commonly understood as a two-stage process involving pre-training and post-training. We point out the need for an additional intermediate stage called reinforcement mid-training with potential for strong performance gains. In this paper, we formally define the problem and identify three key challenges: (1) inefficient training due to excessive reasoning steps, (2) disregard of the imbalanced token entropy distribution, and (3) underutilization of token information. To address these challenges, we propose RMT, a framework for efficient, adaptive, and unified reinforcement mid-training with various innovative components. In particular, we first introduce a dynamic token budget mechanism that constrains unnecessary reasoning steps and mitigates model overthinking. Next, we design a curriculum-based adaptive sampling method that fosters a progressive learning trajectory from easy to hard tokens. Finally, we present a dual training strategy that combines reinforcement learning with next-token prediction, ensuring targeted learning on key tokens and full exploitation of all token information. Extensive experiments demonstrate the superiority of RMT over state-of-the-art methods, achieving up to +64.91% performance improvement with only 21% of the reasoning length in language modeling. We also show that checkpoints obtained after reinforcement mid-training can benefit the subsequent post-training, yielding up to +18.76% improvement in the mathematical domain.

Street view imagery has become an essential source for geospatial data collection and urban analytics, enabling the extraction of valuable insights that support informed decision-making. However, synthesizing street-view images from corresponding satellite imagery presents significant challenges due to substantial differences in appearance and viewing perspective between these two domains. This paper presents a hybrid framework that integrates diffusion-based models and conditional generative adversarial networks to generate geographically consistent street-view images from satellite imagery. Our approach uses a multi-stage training strategy that incorporates Stable Diffusion as the core component within a dual-branch architecture. To enhance the framework's capabilities, we integrate a conditional Generative Adversarial Network (GAN) that enables the generation of geographically consistent panoramic street views. Furthermore, we implement a fusion strategy that leverages the strengths of both models to create robust representations, thereby improving the geometric consistency and visual quality of the generated street-view images. The proposed framework is evaluated on the challenging Cross-View USA (CVUSA) dataset, a standard benchmark for cross-view image synthesis. Experimental results demonstrate that our hybrid approach outperforms diffusion-only methods across multiple evaluation metrics and achieves competitive performance compared to state-of-the-art GAN-based methods. The framework successfully generates realistic and geometrically consistent street-view images while preserving fine-grained local details, including street markings, secondary roads, and atmospheric elements such as clouds.

Deep neural networks remain highly vulnerable to adversarial examples, and most defenses collapse once gradients can be reliably estimated. We identify gradient consensus--the tendency of randomized transformations to yield aligned gradients--as a key driver of adversarial transferability. Attackers exploit this consensus to construct perturbations that remain effective across transformations. We introduce DRIFT (Divergent Response in Filtered Transformations), a stochastic ensemble of lightweight, learnable filters trained to actively disrupt gradient consensus. Unlike prior randomized defenses that rely on gradient masking, DRIFT enforces gradient dissonance by maximizing divergence in Jacobian-and logit-space responses while preserving natural predictions. Our contributions are threefold: (i) we formalize gradient consensus and provide a theoretical analysis linking consensus to transferability; (ii) we propose a consensus-divergence training strategy combining prediction consistency, Jacobian separation, logit-space separation, and adversarial robustness; and (iii) we show that DRIFT achieves substantial robustness gains on ImageNet across CNNs and Vision Transformers, outperforming state-of-the-art preprocessing, adversarial training, and diffusion-based defenses under adaptive white-box, transfer-based, and gradient-free attacks. DRIFT delivers these improvements with negligible runtime and memory cost, establishing gradient divergence as a practical and generalizable principle for adversarial defense. An adaptive adversary can approximate gradients across these transformations (e.g., via Expectation over Transformation (EoT) (Athalye et al., 2018)) and exploit the consistent directions that emerge, leading to transferable adversarial examples. This vulnerability stems not from insufficient randomness, but from the fact that most stochastic defenses still preserve a coherent, low-variance surrogate gradient landscape. We argue that true robustness requires not masking gradients, but destroying their alignment.

In the field of multi-organ medical image segmentation, recent methods frequently employ Transformers to capture long-range dependencies from image features. However, these methods overlook the high computational cost of Transformers and their deficiencies in extracting local detailed information. To address high computational costs and inadequate local detail information, we reassess the design of feature extraction modules and propose a new deep-learning network called LamFormer for fine-grained segmentation tasks across multiple organs. LamFormer is a novel U-shaped network that employs Linear Attention Mamba (LAM) in an enhanced pyramid encoder to capture multi-scale long-range dependencies. We construct the Parallel Hierarchical Feature Aggregation (PHFA) module to aggregate features from different layers of the encoder, narrowing the semantic gap among features while filtering information. Finally, we design the Reduced Transformer (RT), which utilizes a distinct computational approach to globally model up-sampled features. RRT enhances the extraction of detailed local information and improves the network's capability to capture long-range dependencies. LamFormer outperforms existing segmentation methods on seven complex and diverse datasets, demonstrating exceptional performance. Moreover, the proposed network achieves a balance between model performance and model complexity.

Agricultural visual question answering is essential for providing farmers and researchers with accurate and timely knowledge. However, many existing approaches are predominantly developed for evidence-constrained settings such as text-only queries or single-image cases. This design prevents them from coping with real-world agricultural scenarios that often require multi-image inputs with complementary views across spatial scales, and growth stages. Moreover, limited access to up-to-date external agricultural context makes these systems struggle to adapt when evidence is incomplete. In addition, rigid pipelines often lack systematic quality control. To address this gap, we propose a self-reflective and self-improving multi-agent framework that integrates four roles, the Retriever, the Reflector, the Answerer, and the Improver. They collaborate to enable context enrichment, reflective reasoning, answer drafting, and iterative improvement. A Retriever formulates queries and gathers external information, while a Reflector assesses adequacy and triggers sequential reformulation and renewed retrieval. Two Answerers draft candidate responses in parallel to reduce bias. The Improver refines them through iterative checks while ensuring that information from multiple images is effectively aligned and utilized. Experiments on the AgMMU benchmark show that our framework achieves competitive performance on multi-image agricultural QA.

The exponential technological breakthrough of the Fin-Tech industry has significantly enhanced user engagement through sophisticated advisory chatbots. However, large-scale fine-tuning of LLMs can occasionally yield unprofessional or flippant remarks, such as "With that money, you're going to change the world," which, though factually correct, can be contextually inappropriate and erode user trust. The scarcity of domain-specific datasets has led previous studies to focus on isolated components, such as reasoning-aware frameworks or the enhancement of human-like response generation. To address this research gap, we present Fin-Solution 2.O, an advanced solution that 1) introduces the multi-turn financial conversational dataset, Fin-V ault, and 2) incorporates a unified model, Fin-Ally, which integrates commonsense reasoning, politeness, and human-like conversational dynamics. Fin-Ally is powered by COMET -BART -embedded commonsense context and optimized with a Direct Preference Optimization (DPO) mechanism to generate human-aligned responses. The novel Fin-V ault dataset, consisting of 1,417 annotated multi-turn dialogues, enables Fin-Ally to extend beyond basic account management to provide personalized budgeting, real-time expense tracking, and automated financial planning. Our comprehensive results demonstrate that incorporating commonsense context enables language models to generate more refined, textually precise, and professionally grounded financial guidance, positioning this approach as a next-generation AI solution for the FinTech sector.

Multilingual large language models (LLMs) possess impressive multilingual understanding and generation capabilities. However, their performance and cross-lingual alignment often lag for non-dominant languages. A common solution is to fine-tune LLMs on large-scale and more balanced multilingual corpus, but such approaches often lead to imprecise alignment and suboptimal knowledge transfer, struggling with limited improvements across languages. In this paper, we propose AlignX to bridge the multilingual performance gap, which is a two-stage representation-level framework for enhancing multilingual performance of pre-trained LLMs. In the first stage, we align multilingual representations with multilingual semantic alignment and language feature integration. In the second stage, we stimulate the multilingual capability of LLMs via multilingual instruction fine-tuning. Experimental results on several pre-trained LLMs demonstrate that our approach enhances LLMs' multilingual general and cross-lingual generation capability. Further analysis indicates that AlignX brings the multilingual representations closer and improves the cross-lingual alignment.

Advanced deep learning-based approaches have been actively applied to forecast the spatiotemporal physical dynamics governed by partial differential equations (PDEs), which acts as a critical procedure in tackling many science and engineering problems. As real-world physical environments like PDE system parameters are always capricious, how to generalize across unseen out-of-distribution (OOD) forecasting scenarios using limited training data is of great importance. To bridge this barrier, existing methods focus on discovering domain-generalizable representations across various PDE dynamics trajectories. However, their zero-shot OOD generalization capability remains deficient, since extra test-time samples for domain-specific adaptation are still required. This is because the fundamental physical invariance in PDE dynamical systems are yet to be investigated or integrated. To this end, we first explicitly define a two-fold PDE invariance principle, which points out that ingredient operators and their composition relationships remain invariant across different domains and PDE system evolution. Next, to capture this two-fold PDE invariance, we propose a physics-guided invariant learning method termed iMOOE, featuring an Invariance-aligned Mixture Of Operator Expert architecture and a frequency-enriched invariant learning objective. Extensive experiments across simulated benchmarks and real-world applications validate iMOOE's superior in-distribution performance and zero-shot generalization capabilities on diverse OOD forecasting scenarios.