Reinforcement Learning (RL) has proven highly effective for autoregressive language models, but adapting these methods to diffusion large language models (dLLMs) presents fundamental challenges. The core difficulty lies in likelihood approximation: while autoregressive models naturally provide token-level conditional probabilities essential for token-level RL objectives (e.g., GRPO), dLLMs generate sequences through iterative non-autoregressive denoising steps that lack this factorization. To address this fundamental mismatch, we propose ELBO-based Sequence-level Policy Optimization (ESPO), a principled RL framework that treats entire sequence generation as a single action and uses the ELBO as a tractable sequence-level likelihood proxy. Our method incorporates per-token normalization of importance ratios and robust KL-divergence estimation to ensure stable large-scale training. Extensive experiments on mathematical reasoning, coding, and planning tasks demonstrate that ESPO significantly outperforms token-level baselines, achieving dramatic improvements of 20-40 points on the Countdown task, while maintaining consistent gains on math and coding benchmarks. Our approach establishes sequence-level optimization as a principled and empirically effective paradigm for RL in dLLMs. Our code is available at https://github.com/ML-GSAI/ESPO. Large language models (LLMs) (OpenAI, 2023) have become a cornerstone of modern natural language processing, achieving remarkable progress across math (Guo et al., 2025), coding (Hui et al., 2024), and planning tasks (Y ao et al., 2023). While autoregressive (AR) modeling has long dominated this field, recent advances in diffusion large language models (dLLMs) have demonstrated strong potential as an alternative formulation (Ou et al., 2024; Shi et al., 2024; Sahoo et al., 2024; Nie et al., 2025; Y e et al., 2025). With the advent of powerful pretrained dLLMs, the next frontier lies in post-training (Ouyang et al., 2022) to further enhance their capabilities. Among various post-training paradigms, reinforcement learning (RL) has emerged as a powerful approach that enables test-time scaling (Snell et al., 2025) through verifiable rewards (Guo et al., 2025). It has yielded substantial gains on reasoning tasks in recent AR models (OpenAI, 2024), such as math (Cobbe et al., 2021b), coding (Chen et al., 2021), and reasoning (Liu et al., 2023b).

Message Passing Neural Networks (MPNNs) hold a key position in machine learning on graphs, but they struggle with unintended behaviors, such as over-smoothing and over-squashing, due to irregular data structures. The observation and formulation of these limitations have become foundational in constructing more informative graph representations. In this paper, we delve into the limitations of MPNNs, focusing on aspects that have previously been overlooked. Our observations reveal that even within a single layer, the information specific to an individual node can become significantly diluted. To delve into this phenomenon in depth, we present the concept of Over-dilution and formulate it with two dilution factors: intra-node dilution for attribute-level and inter-node dilution for node-level representations. We also introduce a transformer-based solution that alleviates over-dilution and complements existing node embedding methods like MPNNs. Our findings provide new insights and contribute to the development of informative representations. The implementation and supplementary materials are publicly available at https://github.com/LeeJunHyun/NATR.

Reinforcement learning in partially observable environments is typically challenging, as it requires agents to learn an estimate of the underlying system state. These challenges are exacerbated in multi-agent settings, where agents learn simultaneously and influence the underlying state as well as each others' observations. We propose the use of learned beliefs on the underlying state of the system to overcome these challenges and enable reinforcement learning with fully decentralized training and execution. Our approach leverages state information to pre-train a probabilistic belief model in a self-supervised fashion. The resulting belief states, which capture both inferred state information as well as uncertainty over this information, are then used in a state-based reinforcement learning algorithm to create an end-to-end model for cooperative multi-agent reinforcement learning under partial observability. By separating the belief and reinforcement learning tasks, we are able to significantly simplify the policy and value function learning tasks and improve both the convergence speed and the final performance. We evaluate our proposed method on diverse partially observable multi-agent tasks designed to exhibit different variants of partial observability.

Semantic leakage is a phenomenon recently introduced by Gonen et al. (2024). It refers to a situation in which associations learnt from the training data emerge in language model generations in an unexpected and sometimes undesired way. Prior work has focused on leakage in large language models (7B+ parameters). In this study, I use Qwen2.5 model family to explore whether smaller models, ranging from 500M to 7B parameters, demonstrate less semantic leakage due to their limited capacity for capturing complex associations. Building on the previous dataset from Gonen et al. (2024), I introduce a new dataset of color-focused prompts, categorized into specific types of semantic associations, to systematically evaluate the models' performance. Results indicate that smaller models exhibit less semantic leakage overall, although this trend is not strictly linear, with medium-sized models sometimes surpassing larger ones in leaking behavior. The dataset, the model generations, and the evaluation code are publicly available at https://github.com/smilni/semantic_leakage_project.

Graph Neural Networks (GNNs) have revolutionized the domain of graph representation learning by utilizing neighborhood aggregation schemes in many popular architectures, such as message passing graph neural networks (MPGNNs). This scheme involves iteratively calculating a node's representation vector by aggregating and transforming the representation vectors of its adjacent nodes. Despite their effectiveness, MPGNNs face significant issues, such as oversquashing, oversmoothing, and underreaching, which hamper their effectiveness. Additionally, the reliance of MPGNNs on the homophily assumption, where edges typically connect nodes with similar labels and features, limits their performance in heterophilic contexts, where connected nodes often have significant differences. This necessitates the development of models that can operate effectively in both homophilic and heterophilic settings. In this paper, we propose a novel approach, ClassContrast, grounded in Energy Landscape Theory from Chemical Physics, to overcome these limitations. ClassContrast combines spatial and contextual information, leveraging a physics-inspired energy landscape to model node embeddings that are both discriminative and robust across homophilic and heterophilic settings. Our approach introduces contrast-based homophily matrices to enhance the understanding of class interactions and tendencies. Through extensive experiments, we demonstrate that ClassContrast outperforms traditional GNNs in node classification and link prediction tasks, proving its effectiveness and versatility in diverse real-world scenarios.

Knowledge Organization Systems (KOSs), such as term lists, thesauri, taxonomies, and ontologies, play a fundamental role in categorising, managing, and retrieving information. In the academic domain, KOSs are often adopted for representing research areas and their relationships, primarily aiming to classify research articles, academic courses, patents, books, scientific venues, domain experts, grants, software, experiment materials, and several other relevant products and agents. These structured representations of research areas, widely embraced by many academic fields, have proven effective in empowering AI-based systems to i) enhance retrievability of relevant documents, ii) enable advanced analytic solutions to quantify the impact of academic research, and iii) analyse and forecast research dynamics. This paper aims to present a comprehensive survey of the current KOS for academic disciplines. We analysed and compared 45 KOSs according to five main dimensions: scope, structure, curation, usage, and links to other KOSs. Our results reveal a very heterogeneous scenario in terms of scope, scale, quality, and usage, highlighting the need for more integrated solutions for representing research knowledge across academic fields. We conclude by discussing the main challenges and the most promising future directions.

Despite the remarkable ability of large vision-language models (LVLMs) in image comprehension, these models frequently generate plausible yet factually incorrect responses, a phenomenon known as hallucination.Recently, in large language models (LLMs), augmenting LLMs by retrieving information from external knowledge resources has been proven as a promising solution to mitigate hallucinations.However, the retrieval augmentation in LVLM significantly lags behind the widespread applications of LVLM. Moreover, when transferred to augmenting LVLMs, sometimes the hallucination degree of the model is even exacerbated.Motivated by the research gap and counter-intuitive phenomenon, we introduce a novel framework, the Active Retrieval-Augmented large vision-language model (ARA), specifically designed to address hallucinations by incorporating three critical dimensions: (i) dissecting the retrieval targets based on the inherent hierarchical structures of images. (ii) pinpointing the most effective retrieval methods and filtering out the reliable retrieval results. (iii) timing the retrieval process to coincide with episodes of low certainty, while circumventing unnecessary retrieval during periods of high certainty. To assess the capability of our proposed ARA model in reducing hallucination, we employ three widely used LVLM models (LLaVA-1.5, Qwen-VL, and mPLUG-Owl2) across four benchmarks. Our empirical observations suggest that by utilizing fitting retrieval mechanisms and timing the retrieval judiciously, we can effectively mitigate the hallucination problem. We hope that this study can provide deeper insights into how to adapt the retrieval augmentation to LVLMs for reducing hallucinations with more effective retrieval and minimal retrieval occurrences.

The trend of modeless ML inference is increasingly growing in popularity as it hides the complexity of model inference from users and caters to diverse user and application accuracy requirements. Previous work mostly focuses on modeless inference in data centers. To provide low-latency inference, in this paper, we promote modeless inference at the edge. The edge environment introduces additional challenges related to low power consumption, limited device memory, and volatile network environments. To address these challenges, we propose HawkVision, which provides low-latency modeless serving of vision DNNs. HawkVision leverages a two-layer edge-DC architecture that employs confidence scaling to reduce the number of model options while meeting diverse accuracy requirements. It also supports lossy inference under volatile network environments. Our experimental results show that HawkVision outperforms current serving systems by up to 1.6X in P99 latency for providing modeless service. Our FPGA prototype demonstrates similar performance at certain accuracy levels with up to a 3.34X reduction in power consumption.

Existing neural field-based SLAM methods typically employ a single monolithic field as their scene representation. This prevents efficient incorporation of loop closure constraints and limits scalability. To address these shortcomings, we propose a neural mapping framework which anchors lightweight neural fields to the pose graph of a sparse visual SLAM system. Our approach shows the ability to integrate large-scale loop closures, while limiting necessary reintegration. Furthermore, we verify the scalability of our approach by demonstrating successful building-scale mapping taking multiple loop closures into account during the optimization, and show that our method outperforms existing state-of-the-art approaches on large scenes in terms of quality and runtime.