Existing concept customization methods have achieved remarkable outcomes in high-fidelity and multi-concept customization. However, they often neglect the influence on the original model's behavior and capabilities when learning new personalized concepts. To address this issue, we propose PureCC. PureCC introduces a novel decoupled learning objective for concept customization, which combines the implicit guidance of the target concept with the original conditional prediction. This separated form enables PureCC to substantially focus on the original model during training. Moreover, based on this objective, PureCC designs a dual-branch training pipeline that includes a frozen extractor providing purified target concept representations as implicit guidance and a trainable flow model producing the original conditional prediction, jointly achieving pure learning for personalized concepts. Furthermore, PureCC introduces a novel adaptive guidance scale $λ^\star$ to dynamically adjust the guidance strength of the target concept, balancing customization fidelity and model preservation. Extensive experiments show that PureCC achieves state-of-the-art performance in preserving the original behavior and capabilities while enabling high-fidelity concept customization. The code is available at https://github.com/lzc-sg/PureCC.

Existing AI-generated video quality assessment (AIGVQA) methods mainly focus on global perceptual realism and coarse text-video alignment, while overlooking a critical requirement in educational scenarios: concept correctness. In early mathematics education, subtle errors in numerical quantities, geometric relations, or spatial configurations may fundamentally alter the conveyed knowledge despite visually plausible generation. To address this problem, we introduce EduAVQABench, the first benchmark for concept-aware educational AIGV assessment, containing 1,130 videos generated by ten state-of-the-art T2V models together with over 310,650 fine-grained human annotations spanning perceptual quality and semantic alignment. Built upon this benchmark, we further propose EduVQA, a concept-aware AIGVQA framework equipped with a Structured 2D Mixture-of-Experts (S2D-MoE) architecture. By jointly modeling fine-grained concept assessment and overall quality prediction through shared experts and adaptive two-dimensional routing, EduVQA effectively captures subtle concept-level inconsistencies overlooked by conventional global scoring methods. Extensive experiments demonstrate that EduVQA consistently outperforms existing AIGVQA approaches across both perceptual and semantic evaluation tasks while exhibiting strong generalization capability on unseen benchmarks. Code and dataset will be publicly available at: https://github.com/EduVQA/EduVQA.

Over the past years, Automated Essay Scoring (AES) systems have gained increasing attention as scalable and consistent solutions for assessing the proficiency of student writing. Despite recent progress, support for Arabic AES remains limited due to linguistic complexity and scarcity of large publicly-available annotated datasets. In this work, we present Qayyem, a Web-based platform designed to support Arabic AES by providing an integrated workflow for assignment creation, batch essay upload, scoring configuration, and per-trait essay evaluation. Qayyem abstracts the technical complexity of interacting with scoring server APIs, allowing instructors to access advanced scoring services through a user-friendly interface. The platform deploys a number of state-of-the-art Arabic essay scoring models with different effectiveness and efficiency figures.

Significant progress has been made in the field of Instruction-based Image Editing Models (IIEMs). However, while these models demonstrate plausible adherence to instructions and strong reasoning ability on current benchmarks, their ability to edit small objects remains underexplored, despite its importance for precise local editing and refining details in both real and generated images. In this paper, we introduce DeepLookEditBench (DLEBench), the first benchmark dedicated to assessing the abilities of IIEMs in editing small-scale objects. Specifically, we construct a challenging testbed comprising 1889 samples across seven instruction types. In these samples, target objects occupy only 1%-10% of the image area, covering complex scenarios such as partial occlusion and multi-object editing. To ensure robust evaluation on this benchmark, we propose an evaluation protocol with refined score rubrics to minimize subjectivity and ambiguity in two criteria: Instruction Following and Visual Consistency. This protocol also introduces a dual-mode evaluation framework (Tool-driven and Oracle-guided Modes) addressing the misalignment between LMM-as-a-Judge and human judgements on DLEBench. Empirical results on 10 IIEMs reveal significant performance gaps in small-scale object editing, highlighting the need for specialized benchmarks to advance this ability.

Reinforcement learning (RL) has equipped LLM agents with a strong ability to solve complex tasks. However, existing RL methods normally use a \emph{single} policy network, causing \emph{simplicity bias} where simple tasks occupy most parameters and dominate gradient updates, leaving insufficient capacity for complex tasks. A plausible remedy could be employing the Mixture-of-Experts (MoE) architecture in the policy network, as MoE allows different parameters (experts) to specialize in different tasks, preventing simple tasks from dominating all parameters. However, a key limitation of traditional MoE is its token-level routing, where the router assigns each token to specialized experts, which fragments phase-consistent patterns into scattered expert assignments and thus undermines expert specialization. In this paper, we propose \textbf{Phase-Aware Mixture of Experts (PA-MoE)}. It first features a lightweight \emph{phase router} that learns latent phase boundaries directly from the RL objective without pre-defining phase categories. Then, the phase router allocates temporally consistent assignments to the same expert, allowing experts to preserve phase-specific expertise. Experimental results demonstrate the effectiveness of our proposed PA-MoE.

On two-sided matching platforms such as online dating and recruiting, recommendation algorithms often aim to maximize the total number of matches. However, this objective creates an imbalance, where some users receive far too many matches while many others receive very few and eventually abandon the platform. Retaining users is crucial for many platforms, such as those that depend heavily on subscriptions. Some may use fairness objectives to solve the problem of match maximization. However, fairness in itself is not the ultimate objective for many platforms, as users do not suddenly reward the platform simply because exposure is equalized. In practice, where user retention is often the ultimate goal, casually relying on fairness will leave the optimization of retention up to luck. In this work, instead of maximizing matches or axiomatically defining fairness, we formally define the new problem setting of maximizing user retention in two-sided matching platforms. To this end, we introduce a dynamic learning-to-rank (LTR) algorithm called Matching for Retention (MRet). Unlike conventional algorithms for two-sided matching, our approach models user retention by learning personalized retention curves from each user's profile and interaction history. Based on these curves, MRet dynamically adapts recommendations by jointly considering the retention gains of both the user receiving recommendations and those who are being recommended, so that limited matching opportunities can be allocated where they most improve overall retention. Naturally but importantly, empirical evaluations on synthetic and real-world datasets from a major online dating platform show that MRet achieves higher user retention, since conventional methods optimize matches or fairness rather than retention.

The ability of machine learning models to store input information in hidden layer vector embeddings, analogous to the concept of `memory', is widely employed but not well characterized. We find that language model embeddings typically contain relatively little input information regardless of data and compute scale during training. In contrast, embeddings from autoencoders trained for input regeneration are capable of nearly perfect memory formation. The substitution of memory embeddings for token sequences leads to substantial computational efficiencies, motivating the introduction of a parallelizable encoder-decoder memory model architecture. Upon causal training these models contain information-poor embeddings incapable of arbitrary information access, but by combining causal and information retention objective functions they learn to form and decode information-rich memories. Training can be further streamlined by freezing a high fidelity encoder followed by a curriculum training approach where decoders first learn to process memories and then learn to additionally predict next tokens. We introduce the perspective that next token prediction training alone is poorly suited for accurate memory formation as the objective itself is non-invertible, motivating the use of combined objective functions for models where the entire input is not exposed.

Audio diffusion models can synthesize high-fidelity music from text, yet achieving fine-grained control over specific musical attributes remains challenging, as their internal mechanisms for representing high-level concepts are poorly understood. In this work, we use activation patching to demonstrate that recent audio diffusion architectures exhibit a semantic bottleneck, where a small, shared subset of consecutive attention layers controls distinct musical concepts, such as the presence of specific instruments, vocals, or genres. Building on this, we systematically evaluate a broad spectrum of steering paradigms, comparing activation steering against prompt-level, score-space, and weight-space interventions, analyzing the interaction between the steering mechanism and the intervention site. Our new benchmark, supported by an extensive user study, demonstrates that localized activation steering establishes a new state-of-the-art in audio concept modulation.

Advances in large language models (LLMs) are driving a shift toward using reinforcement learning (RL) to train agents from iterative, multi-turn interactions across tasks. However, multi-turn RL remains challenging as rewards are often sparse or delayed, and environments can be stochastic. In this regime, naive trajectory sampling can hinder exploitation and induce mode collapse. We propose TSR (Trajectory-Search Rollouts), a training-time approach that repurposes test-time scaling ideas for improved per-turn rollout generation. TSR performs lightweight tree-style search to construct high-quality trajectories by selecting high-scoring actions at each turn using state-based feedback. This improves rollout quality and stabilizes learning while remaining compatible with standard policy gradient optimizers, making TSR optimizer-agnostic. We instantiate TSR with best-of-N, beam, and shallow lookahead search, and pair it with PPO and GRPO, achieving up to 15% performance gains and more stable learning on Sokoban, FrozenLake, and WebShop tasks at a modest, one-time increase in training compute. By moving search from inference time to the rollout stage of training, TSR provides a modular and general mechanism for stronger multi-turn agent learning, complementary to existing frameworks and rejection-sampling-style selection methods.

Human activity recognition (HAR) on resource constrained devices requires high accuracy across diverse sensor setups. Selective state space models (SSMs) offer efficient linear time sequence processing, presenting a compelling alternative to attention mechanisms. However, their TinyML design space remains unexplored. This paper introduces BabyMamba-HAR, comprising two lightweight architectures: (1) CI-BabyMamba-HAR, utilizing a channel independent stem for noise robustness, and (2) Crossover-BiDir-BabyMamba-HAR, utilizing an early fusion stem for channel count independent complexity. Both integrate weight tied bidirectional scanning and gated temporal attention pooling. Across eight benchmarks, Crossover-BiDir-BabyMamba-HAR averages an 86.52% F1-score with 27K parameters and 2.21M MACs, matching TinyHAR (86.16%) while requiring 11x fewer MACs on high channel datasets. On-device deployment on the Raspberry Pi Pico 2 and ESP32 utilized a mixed precision C++ runtime (INT8 projections, float32 states). A fused computation strategy with lifetime aware memory management reduces peak memory footprint from O(B*dmodel*L*dstate) to O(B*dmodel*dstate), adapting to support weight-tied bidirectional and channel-streaming execution. Both architectures achieved full 8/8 dataset coverage with >99.2% PyTorch parity, whereas INT8 quantized TFLite baselines showed degraded coverage and parity (TinyHAR: 7/8 and 4/8 coverage at 60.4% and 88.6% parity, TinierHAR: 8/8 and 6/8 at 54.2% and 90.8%, DeepConvLSTM: 1/8 and 0/8 on Pico 2 and ESP32, respectively). Crossover-BiDir-BabyMamba-HAR averages 154.4 ms latency on ESP32 and 481.9 ms on Pico 2. Ablations confirm bidirectional scanning and gated attention improve F1-scores by up to 8.42% and 8.94%, respectively, establishing practical principles for TinyML SSM deployment.

In robot-assisted minimally invasive surgery (RMIS), reduced haptic feedback and depth cues increase reliance on expert visual perception, motivating gaze-guided training and learning-based surgical perception models. However, operative expert gaze is costly to collect, and it remains unclear how the source of gaze supervision, both expertise level (intermediate vs. novice) and perceptual modality (active execution vs. passive viewing), shapes what attention models learn. We introduce a paired active-passive, multi-task surgical gaze dataset collected on the da Vinci SimNow simulator across four drills. Active gaze was recorded during task execution using a VR headset with eye tracking, and the corresponding videos were reused as stimuli to collect passive gaze from observers, enabling controlled same-video comparisons. We quantify skill- and modality-dependent differences in gaze organization and evaluate the substitutability of passive gaze for operative supervision using fixation density overlap analyses and single-frame saliency modeling. Across settings, MSI-Net produced stable, interpretable predictions, whereas SalGAN was unstable and often poorly aligned with human fixations. Models trained on passive gaze recovered a substantial portion of intermediate active attention, but with predictable degradation, and transfer was asymmetric between active and passive targets. Notably, novice passive labels approximated intermediate-passive targets with limited loss on higher-quality demonstrations, suggesting a practical path for scalable, crowd-sourced gaze supervision in surgical coaching and perception modeling.

Despite strong generalization capabilities, Vision-Language-Action (VLA) models remain constrained by the high cost of expert demonstrations and limited real-world interaction. While online reinforcement learning (RL) has shown promise, its application to real-world VLA manipulation is hindered by low exploration efficiency and restricted exploration coverage. Through systematic real-world experiments, we observe that the effective exploration space of online RL is largely constrained by the trajectory distribution induced during supervised fine-tuning (SFT). Motivated by this observation, we propose TwinRL, a digital twin-real-world collaborative post-training framework that expands and guides RL exploration for VLA models through three stages: SFT warm-up, twin RL warm-up, and real-world RL. TwinRL first reconstructs a high-fidelity digital twin from smartphone-captured scenes. During the SFT stage, we introduce an exploration space expansion strategy that expands the support of the trajectory distribution beyond real demonstrations, reshaping the exploration space for more effective RL. Rather than treating the twin as a data augmentation tool, we propose a twin RL warm-up strategy that enables it to act as an exploration guide for real-world RL. Specifically, TwinRL performs efficient parallel RL in the digital twin to generate interactive trajectories that populate the replay buffer and stabilize subsequent real-world RL learning. This process also identifies failure-prone yet informative configurations, enabling targeted human-in-the-loop rollouts to further improve on-robot efficiency. Across four tasks, TwinRL achieves near-100% success in both in-distribution and out-of-distribution regions, delivering over 30% faster convergence than prior real-world RL methods with only 20 minutes of on-robot interaction.

Reliable models should not only predict correctly, but also justify decisions with acceptable evidence. Yet conventional supervised learning typically provides only class-level labels, allowing models to achieve high accuracy through shortcut correlations rather than the intended evidence. Human priors can help constrain such behavior, but aligning models to these priors remains challenging because learned representations often diverge from human perception. To address this challenge, we propose an attribution-based human prior alignment method. We encode human priors as input regions that the model is expected to rely on (e.g., bounding boxes), and leverage a highly faithful subset-selection-based attribution approach to expose the model's decision evidence during training. When the attribution region deviates substantially from the prior regions, we penalize reliance on off-prior evidence, encouraging the model to shift its attribution toward the intended regions. This is achieved through a training objective that imposes attribution constraints induced by the human prior. We validate our method on both image classification and click decision tasks in MLLM-based GUI agent models. Across conventional classification and autoregressive generation settings, human prior alignment consistently improves task accuracy while also enhancing the model's decision reasonability.

Masked diffusion language models generate text through iterative masked-token filling, but terminal-only rewards on final completions provide coarse credit assignment for the intermediate filling decisions that shape the generation process. We propose Diffusion-State Policy Optimization (DiSPO), a plug-in credit-assignment layer that directly optimizes intermediate filling decisions. At selected intermediate masked states, DiSPO branches by resampling the currently masked positions from rollout-cached logits, scores the resulting completions, and updates only the newly filled tokens, requiring no additional multi-step diffusion rollouts or optimizer steps. We formalize a fixed-state objective for branched completions and derive a policy-gradient estimator that reuses the same rollouts as terminal-feedback policy optimization. Experiments on LLaDA-8B-Instruct show that DiSPO consistently improves terminal-feedback baselines, including diffu-GRPO and SPG, on math and planning benchmarks under matched rollout compute and optimizer steps, supporting its use as a general plug-in for masked diffusion policy optimization. Our project page is available at https://daioba.github.io/dispo .

Low-rank adaptation (LoRA) approximates the update of a pretrained weight matrix using the product of two low-rank matrices. However, standard LoRA follows an explicit-rank paradigm, where increasing model capacity requires adding more rows or columns (i.e., basis vectors) to the low-rank matrices, leading to substantial parameter growth. In this paper, we find that these basis vectors exhibit significant parameter redundancy and can be compactly represented by lightweight nonlinear functions. Therefore, we propose Generative Low-Rank Adapter (GenLoRA), which replaces explicit basis vector storage with nonlinear basis vector generation. Specifically, GenLoRA maintains a latent vector for each low-rank matrix and employs a set of lightweight radial basis functions (RBFs) to synthesize the basis vectors. Each RBF requires far fewer parameters than an explicit basis vector, enabling higher parameter efficiency in GenLoRA. Extensive experiments across multiple datasets and architectures show that GenLoRA attains higher effective LoRA ranks under smaller parameter budgets, resulting in superior fine-tuning performance. The code is available at https://anonymous.4open.science/r/GenLoRA.

Low-Rank Adaptation (LoRA) is the prevailing approach for efficient large language model (LLM) fine-tuning. Building on this paradigm, recent studies have proposed alternative initialization strategies, architectural modifications, and optimization adjustments, reporting substantial improvements over vanilla LoRA. However, these gains are often demonstrated under fixed or narrowly tuned hyperparameter settings, despite the known sensitivity of neural networks to training configurations. In this work, we systematically re-evaluate nine representative LoRA variants alongside vanilla LoRA through extensive hyperparameter searches over learning rate, batch size, rank, and training duration. Across tasks spanning mathematical reasoning, commonsense reasoning, code generation, and instruction following at diverse model scales, we find that different LoRA methods favor distinct learning rate ranges. Crucially, once learning rates are properly tuned, all methods achieve similar peak performance (within 1-2%), with only subtle rank-dependent behaviors. These results suggest that vanilla LoRA remains a competitive baseline and that improvements reported under a single training configuration may not reflect consistent methodological advantages. Finally, a second-order analysis attributes the differing optimal learning rate ranges to variations in the largest Hessian eigenvalue, aligning with classical learning theories.

Reinforcement learning has been widely applied to diffusion and flow models for visual tasks such as text-to-image generation. However, these tasks remain challenging because diffusion models have intractable likelihoods, which creates a barrier for directly applying popular policy-gradient type methods. Existing approaches primarily focus on crafting new objectives built on already heavily engineered LLM objectives, using ad hoc estimators for likelihood, without a thorough investigation into how such estimation affects overall algorithmic performance. In this work, we provide a systematic analysis of the RL design space by disentangling three factors: i) policy-gradient objectives, ii) likelihood estimators, and iii) rollout sampling schemes. We show that adopting an evidence lower bound (ELBO) based model likelihood estimator, computed only from the final generated sample, is the dominant factor enabling effective, efficient, and stable RL optimization, outweighing the impact of the specific policy-gradient loss functional. We validate our findings across multiple reward benchmarks using SD 3.5 Medium, and observe consistent trends across all tasks. Our method improves the GenEval score from 0.24 to 0.95 in 90 GPU hours, which is $4.6\times$ more efficient than FlowGRPO and $2\times$ more efficient than the SOTA method DiffusionNFT without reward hacking.

Learning from a stream of tasks usually pits plasticity against stability: acquiring new knowledge often causes catastrophic forgetting of past information. Most methods address this by summing competing loss terms, creating gradient conflicts that are managed with complex and often inefficient strategies such as external memory replay or parameter regularization. We propose a reformulation of the continual learning objective using Douglas-Rachford Splitting (DRS). This reframes the learning process not as a direct trade-off, but as a negotiation between two decoupled objectives: one promoting plasticity for new tasks and the other enforcing stability of old knowledge. By iteratively finding a consensus through their proximal operators, DRS provides a more principled and stable learning dynamic. Our approach achieves an efficient balance between stability and plasticity without the need for auxiliary modules or complex add-ons, providing a simpler yet more powerful paradigm for continual learning systems.

Real-time polyp segmentation is essential for early colorectal cancer detection, yet clinical deployment remains blocked by GPU dependency. We introduce the UltraSeg family, a set of CPU-native segmentation models operating below 0.3M parameters. UltraSeg-108K (0.108M) establishes the extreme-compression frontier, while UltraSeg-130K (0.130M) integrates cross-layer lightweight fusion for enhanced multi-center generalization. The architecture replaces parameter-heavy components with grouped multi-rate dilated convolutions and attention-gated cross-layer fusion, achieving real-time throughput on a single CPU core (exceeding 50 FPS at 256*256 and 30 FPS at 352*352) without sacrificing clinical-grade accuracy. Evaluated on seven public datasets, UltraSeg-130K attains Dice scores exceeding 0.8 at both resolutions, substantially outperforming all existing sub-0.3M competitors. Notably, it approaches or exceeds UNet-Medium (7.76M parameters) on zero-shot external validations while using only 1.7% of its parameters, establishing the first strong baseline for CPU-native real-time polyp segmentation. When scaled to 4.38M parameters, UltraSeg achieves accuracy competitive with heavyweight state-of-the-art models while maintaining an order-of-magnitude parameter advantage, demonstrating that the proposed design principles yield intrinsic representational gains across the entire efficiency spectrum. By delivering the first clinically deployable, CPU-native real-time solution, this work provides an immediately usable tool for resource-limited settings and a reproducible blueprint for real-time medical AI beyond endoscopy. Source code is publicly available.

Despite recent progress in vision-language models (VLMs), existing approaches often fail to generate personalized responses based on the user's specific experiences, as they lack the ability to associate visual inputs with a user's accumulated visual-textual context. We newly formalize this challenge as contextualized visual personalization, which requires the visual recognition and textual retrieval of personalized visual experiences by VLMs when interpreting new images. To address this issue, we propose CoViP, a unified framework that treats personalized image captioning as a core task for contextualized visual personalization and improves this capability through reinforcement-learning-based post-training and caption-augmented generation. We further introduce diagnostic evaluations that explicitly rule out textual shortcut solutions and verify whether VLMs truly leverage visual context. Extensive experiments demonstrate that existing open-source and proprietary VLMs exhibit substantial limitations, while CoViP not only improves personalized image captioning but also yields holistic gains across downstream personalization tasks. These results highlight CoViP as a crucial stage for enabling robust and generalizable contextualized visual personalization.

Distribution matching distillation (DMD) facilitates few-step image generation by aligning a distilled student with a reference multi-step teacher. In practice, however, optimizing DMD can reduce sample diversity in few-step synthesis, and existing remedies typically rely on perceptual or adversarial regularization, leading to stability and scalability challenges during training. Here, we describe diversity-preserved DMD (DP-DMD), a role-separated distillation method inspired by the complementary roles of early and late denoising steps. Specifically, the first distillation step is trained with a teacher-derived target-prediction objective (e.g., v-prediction) to preserve sample diversity, while the remaining steps are optimized with the standard DMD loss to refine perceptual quality. DP-DMD, with no perceptual or adversarial regularization, no additional modules, and no teacher-generated reference samples, preserves sample diversity while maintaining competitive visual quality under few-step sampling, providing a simple and stable alternative to other DMD variants.

Artificial intelligence has shown remarkable success in materials discovery and property prediction, particularly for crystalline and polymer systems where material properties and structures are dominated by discrete graph representations. Such graph-central paradigm breaks down on composite materials, which possess continuous and nonlinear design spaces. General composite descriptors, e.g., fiber volume and misalignment angle, cannot fully capture the fiber distributions that determine microstructural characteristics, necessitating the integration of heterogeneous data sources through multimodal learning. Existing alignment-oriented frameworks have proven effective on abundant crystal or polymer data under discrete, unique graph-property mapping assumptions, but fail to address the highly continuous composite design space under extreme data scarcity. In this work we introduce ORDinal-aware imagE-tabulaR alignment (ORDER), a multimodal pretraining framework that establishes ordinality as a core principle for material representations. ORDER ensures that materials with similar target properties occupy nearby regions in the latent space, which effectively preserves the continuous nature of composite properties and enables meaningful interpolation between sparsely observed designs. We evaluate ORDER on a Nanofiber-reinforced composite dataset and a carbon fiber T700 dataset. ORDER and its variants outperform both alignment-oriented and customized property-aware contrastive baselines across property prediction, cross-modal retrieval, and microstructure generation tasks. We further introduce physics-based ordinal surrogate signals avoiding the need for full property annotation during pretrain. Our work demonstrates learning continuous multimodal features are fundamental for composite materials, and provides a reliable pathway toward data-efficient universal multimodal intelligent systems.

Group Relative Policy Optimization (GRPO) has become the dominant method for reinforcement learning with verifiable rewards in large language models, but it suffers from two critical limitations: gradient vanishing and diversity collapse. When training questions are too easy or too hard, all sampled responses receive identical rewards, yielding zero gradients. Meanwhile, the model tends to collapse its responses toward a single reasoning pattern rather than exploring diverse strategies. We propose Transformation-Augmented GRPO (TA-GRPO), a simple but effective method that addresses both issues via question rephrasing. For each training question, we automatically generate multiple problem-equivalent rephrasings that alter wording, format, and information order while preserving the underlying meaning. Because these rephrasings shift the model's perceived difficulty, pooling responses across the original and its rephrasings yields mixed rewards and more diverse reasoning paths. TA-GRPO jointly computes advantages over this expanded response set and aligns all importance ratios to the original question, enabling the model to learn from a richer set of solution attempts. Experiments on four LLMs (Qwen3-1.7B, Qwen3-4B, Llama-3.2-1B, Llama-3.2-3B) show that TA-GRPO consistently improves pass@$k$ on competition-level benchmarks (AMC, OlympiadBench, AIME24, AIME25) and out-of-distribution benchmarks (Minerva, GPQA-Diamond). Notably, it improves the average pass@32 of Qwen3-1.7B and Qwen3-4B by \textbf{4.97} and \textbf{4.34} points, respectively, and matches the exploration quality of baselines trained on up to 2.5$\times$ more data.

Reinforcement Learning (RL) post-training alignment for language models is effective, but also costly and unstable in practice, owing to its complicated training process. To address this, we propose a training-free inference method to sample directly from the optimal RL policy. The transition probability applied to Masked Language Modeling (MLM) consists of a reference policy model and an energy term. Based on this, our algorithm, Energy-Guided Test-Time Scaling (ETS), estimates the key energy term via online Monte Carlo, with a provable convergence rate. Moreover, to ensure practical efficiency, ETS leverages modern acceleration frameworks alongside tailored importance sampling estimators, substantially reducing inference latency while provably preserving sampling quality. Experiments on MLM (including autoregressive models and diffusion language models) across reasoning, coding, and science benchmarks show that our ETS consistently improves generation quality, validating its effectiveness and design. The code is available at https://github.com/sheriyuo/ETS.

Diffusion models have demonstrated exceptional success in video super-resolution (VSR), exhibiting powerful capabilities for generating fine-grained details. However, their potential for space-time video super-resolution (STVSR), which necessitates not only recovering realistic high-resolution visual content but also improving the frame rate with coherent temporal dynamics, remains largely underexplored. Moreover, existing STVSR methods predominantly address spatiotemporal upsampling under simple degradation assumptions, thus failing in real-world scenarios with complex unknown degradations. To address these challenges, we propose OSDEnhancer, the first framework that achieves robust STVSR in one-step diffusion. OSDEnhancer begins with a linear initialization to establish essential spatiotemporal structures and adapt the model for one-step reconstruction. It then applies a divide-and-conquer strategy, introducing the temporal coherence (TC) and texture enrichment (TE) LoRAs that progressively specialize in inter-frame dynamics modeling and fine-grained texture recovery, respectively, while collaborating during inference for enhanced overall performance. A bidirectional VAE decoder employs deformable recurrent blocks to leverage the multi-scale structure of the vanilla VAE, enhancing latent-to-pixel reconstruction through joint multi-scale deformable aggregation and inter-frame feature propagation. Experimental results demonstrate that the proposed method attains state-of-the-art performance with superior generalization in real-world scenarios. The code is available at https://github.com/W-Shuoyan/OSDEnhancer.

Camera redirection aims to replay a dynamic scene from a single monocular video under a user-specified camera trajectory. However, large-angle redirection is inherently ill-posed: a monocular video captures only a narrow spatio-temporal view of a dynamic 3D scene, providing severely limited observations of the underlying 4D world. The key challenge is therefore to recover a complete and coherent representation from this limited input, with consistent geometry and motion. While recent diffusion-based methods achieve impressive visual generation quality, they often break down under large-angle viewpoint changes far from the original trajectory, where missing visual grounding leads to severe geometric ambiguity and temporal inconsistency. We present FreeOrbit4D, an effective training-free framework that tackles this ambiguity by recovering a foreground-complete 4D proxy as structural grounding for video generation. We obtain this proxy by decoupling foreground and background reconstructions: we unproject the monocular video into a static background and partial foreground point clouds in a unified global space, then use an object-centric multi-view diffusion model to synthesize multi-view images and reconstruct complete foreground point clouds in canonical object space. By aligning the canonical foreground point cloud to the global scene space via dense pixel-synchronized 3D-3D correspondences and projecting the foreground-complete 4D proxy onto target camera viewpoints, we provide geometric scaffolds that guide a conditional video diffusion model. Extensive experiments show that FreeOrbit4D produces more faithful and temporally coherent redirected videos under challenging large-angle trajectories, and our proxy further enables applications such as edit propagation and 4D data generation. Project page: https://freeorbit4d.vision.ischool.illinois.edu/

Large Language Models (LLMs) exhibit remarkable capabilities, yet it remains unclear to what extent these reflect sophisticated recall or genuine reasoning ability. We introduce chess as a controlled testbed aimed at disentangling these faculties. Leveraging the game's structure and scalable engine evaluations, we construct a taxonomy of positions varying in density of relevant priors - ranging from common states solvable by memorization to completely novel ones requiring generalization. Crucially, our approach achieves this distinction without requiring explicit knowledge of the models' training data. Applying this taxonomy, we combine a longitudinal analysis of the GPT lineage with a rigorous evaluation of contemporary models, including Claude Opus and Gemini. Our analysis reveals a steep gradient: performance consistently degrades as the density of relevant priors decreases. Notably, for tasks with few relevant priors, base model performance regresses to the random-play baseline. While newer models improve, progress slows significantly for tasks with sparse priors. Furthermore, while reasoning-augmented inference improves performance, its relative marginal benefit per token decreases in the absence of relevant priors. These results suggest limitations in systematic generalization, highlighting the need for mechanisms beyond scale to achieve robust performance when deprived of relevant priors.

Melanoma detection is vital for early diagnosis and effective treatment. While deep learning models on dermoscopic images have shown promise, they require specialized equipment, limiting their use in broader clinical settings. This study introduces a multi-modal melanoma detection system using conventional photo images, making it more accessible and versatile. Our system integrates image data with tabular metadata, such as patient demographics and lesion characteristics, to improve detection accuracy. It employs a multi-modal neural network combining image and metadata processing and supports a two-step model for cases with or without metadata. A three-stage pipeline further refines predictions by boosting algorithms and enhancing performance. To address the challenges of a highly imbalanced dataset, specific techniques were implemented to ensure robust training. An ablation study evaluated recent vision architectures, boosting algorithms, and loss functions, achieving a peak Partial ROC AUC of 0.18068 (0.2 maximum) and top-15 retrieval sensitivity of 0.78371. Results demonstrate that integrating photo images with metadata in a structured, multi-stage pipeline yields significant performance improvements. This system advances melanoma detection by providing a scalable, equipment-independent solution suitable for diverse healthcare environments, bridging the gap between specialized and general clinical practices.

We propose Q-learning with Adjoint Matching (QAM), a novel TD-based reinforcement learning (RL) algorithm that tackles a long-standing challenge in continuous-action RL: efficient optimization of an expressive diffusion or flow-matching policy with respect to a parameterized Q-function. Effective optimization requires exploiting the first-order information of the critic, but it is challenging to do so for flow or diffusion policies because direct gradient-based optimization via backpropagation through their multi-step denoising process is numerically unstable. Existing methods work around this either by only using the value and discarding the gradient information, or by relying on approximations that sacrifice policy expressivity or bias the learned policy. QAM sidesteps both of these challenges by leveraging adjoint matching, a recently proposed technique in generative modeling, which transforms the critic's action gradient to form a step-wise objective function that is free from unstable backpropagation, while providing an unbiased, expressive policy at the optimum. Combined with temporal-difference backup for critic learning, QAM consistently outperforms prior approaches on hard, sparse reward tasks in both offline and offline-to-online RL.

Estimating the unknown reward functions driving agents' behaviors is of central interest in inverse reinforcement learning and game theory. To tackle this problem, we develop a unified framework for reward function recovery in two-player zero-sum matrix games and Markov games with entropy regularization, where we aim to reconstruct the underlying reward functions given observed players' strategies and actions. This task is challenging due to the inherent ambiguity of inverse problems, the non-uniqueness of feasible rewards, and limited observational data coverage. To address these challenges, we establish the reward function's identifiability using the quantal response equilibrium (QRE) under linear assumptions. Building upon this theoretical foundation, we propose a novel algorithm to learn reward functions from observed actions. Our algorithm works in both static and dynamic settings and is adaptable to incorporate different methods, such as Maximum Likelihood Estimation (MLE). We provide strong theoretical guarantees for the reliability and sample efficiency of our algorithm. Further, we conduct extensive numerical studies to demonstrate the practical effectiveness of the proposed framework, offering new insights into decision-making in competitive environments.