ConsRoute:Consistency-Aware Adaptive Query Routing for Cloud-Edge-Device Large Language Models

Large language models (LLMs) deliver impressive capabilities but incur substantial inference latency and cost, which hinders their deployment in latency-sensitive and resource-constrained scenarios. Cloud-edge-device collaborative inference has emerged as a promising paradigm by dynamically routing queries to models of different capacities across tiers. In this paper, we propose ConsRoute, a lightweight, semantic-aware, and adaptive routing framework that significantly improves inference efficiency while minimizing impact on response quality. Unlike prior routing methods that rely on predicting coarse-grained output quality gaps, ConsRoute leverages a reranker to directly assess the semantic consistency between responses generated by models at different tiers, yielding fine-grained soft supervision signals for routing. To minimize device-side overhead, ConsRoute reuses hidden states from the LLM prefilling stage as compact query representations, avoiding additional encoders or inference passes. Furthermore, these representations are clustered, and Bayesian optimization is employed to learn cluster-specific routing thresholds that dynamically balance quality, latency, and cost under heterogeneous query distributions. Extensive experiments demonstrate that ConsRoute achieves near-cloud performance (>=95%) while reducing end-to-end latency and inference cost by nearly 40%, consistently outperforming existing routing baselines in both response quality and system efficiency.

ProRL Agent: Rollout-as-a-Service for RL Training of Multi-Turn LLM Agents

Multi-turn LLM agents are increasingly important for solving complex, interactive tasks, and reinforcement learning (RL) is a key ingredient for improving their long-horizon behavior. However, RL training requires generating large numbers of sandboxed rollout trajectories, and existing infrastructures often couple rollout orchestration with the training loop, making systems hard to migrate and maintain. Under the rollout-as-a-service philosophy, we present ProRL Agent , a scalable infrastructure that serves the full agentic rollout lifecycle through an API service. ProRL Agent also provides standardized and extensible sandbox environments that support diverse agentic tasks in rootless HPC settings. We validate ProRL Agent through RL training on software engineering, math, STEM, and coding tasks. ProRL Agent is open-sourced and integrated as part of NVIDIA NeMo Gym.

Internalizing Agency from Reflective Experience

Large language models are increasingly deployed as autonomous agents that must plan, act, and recover from mistakes through long-horizon interaction with environments that provide rich feedback. However, prevailing outcome-driven post-training methods (e.g., RL with verifiable rewards) primarily optimize final success signals, leaving rich environment feedback underutilized. Consequently, they often lead to distribution sharpening: the policy becomes better at reproducing a narrow set of already-successful behaviors, while failing to improve the feedback-grounded agency needed to expand problem-solving capacity (e.g., Pass@k) in long-horizon settings. To address this, we propose LEAFE (Learning Feedback-Grounded Agency from Reflective Experience), a framework that internalizes recovery agency from reflective experience. Specifically, during exploration, the agent summarizes environment feedback into actionable experience, backtracks to earlier decision points, and explores alternative branches with revised actions. We then distill these experience-guided corrections into the model through supervised fine-tuning, enabling the policy to recover more effectively in future interactions. Across a diverse set of interactive coding and agentic tasks under fixed interaction budgets, LEAFE consistently improves Pass@1 over the base model and achieves higher Pass@k than outcome-driven baselines (GRPO) and experience-based methods such as Early Experience, with gains of up to 14% on Pass@128.

Bridging Sequential and Contextual Features with a Dual-View of Fine-grained Core-Behaviors and Global Interest-Distribution

Click-through rate (CTR) prediction tasks typically estimate the probability of a user clicking on a candidate item by modeling both user behavior sequence features and the item's contextual features, where the user behavior sequence is particularly critical as it dynamically reflects real-time shifts in user interest. Traditional CTR models often aggregate this dynamic sequence into a single vector before interacting it with contextual features. This approach, however, not only leads to behavior information loss during aggregation but also severely limits the model's capacity to capture interactions between contextual features and specific user behaviors, ultimately impairing its ability to capture fine-grained behavioral details and hindering models' prediction accuracy. Conversely, a naive approach of directly interacting with each user action with contextual features is computationally expensive and introduces significant noise from behaviors irrelevant to the candidate item. This noise tends to overwhelm the valuable signals arising from interactions involving more behaviors relevant to the candidate item. Therefore, to resolve the above issue, we propose a Core-Behaviors and Distributional-Compensation Dual-View Interaction Network (CDNet), which bridges the gap between sequential and contextual feature interactions from two complementary angles: a fine-grained interaction involving the most relevant behaviors and contextual features, and a coarse-grained interaction that models the user's overall interest distribution against the contextual features. By simultaneously capturing important behavioral details without forgoing the holistic user interest, CDNet effectively models the interplay between sequential and contextual features without imposing a significant computational burden. Ultimately, extensive experiments validate the effectiveness of CDNet.

When Drafts Evolve: Speculative Decoding Meets Online Learning

Speculative decoding has emerged as a widely adopted paradigm for accelerating large language model inference, where a lightweight draft model rapidly generates candidate tokens that are then verified in parallel by a larger target model. However, due to limited model capacity, drafts often struggle to approximate the target distribution, resulting in shorter acceptance lengths and diminished speedup. A key yet under-explored observation is that speculative decoding inherently provides verification feedback that quantifies the deviation between the draft and target models at no additional cost. This process naturally forms an iterative "draft commits-feedback provides-draft adapts" evolving loop, which precisely matches the online learning paradigm. Motivated by this connection, we propose OnlineSpec, a unified framework that systematically leverages interactive feedback to continuously evolve draft models. Grounded in dynamic regret minimization, we establish a formal link between online learning performance and speculative system's acceleration rate, and develop novel algorithms via modern online learning techniques, including optimistic online learning that adaptively reuses historical gradients as predictive update hints, and online ensemble learning that dynamically maintains multiple draft models. Our algorithms are equipped with theoretical justifications and improved acceleration rates, achieving up to 24% speedup over seven benchmarks and three foundation models.

Strategic Navigation or Stochastic Search? How Agents and Humans Reason Over Document Collections

Multimodal agents offer a promising path to automating complex document-intensive workflows. Yet, a critical question remains: do these agents demonstrate genuine strategic reasoning, or merely stochastic trial-and-error search? To address this, we introduce MADQA, a benchmark of 2,250 human-authored questions grounded in 800 heterogeneous PDF documents. Guided by Classical Test Theory, we design it to maximize discriminative power across varying levels of agentic abilities. To evaluate agentic behaviour, we introduce a novel evaluation protocol measuring the accuracy-effort trade-off. Using this framework, we show that while the best agents can match human searchers in raw accuracy, they succeed on largely different questions and rely on brute-force search to compensate for weak strategic planning. They fail to close the nearly 20% gap to oracle performance, persisting in unproductive loops. We release the dataset and evaluation harness to help facilitate the transition from brute-force retrieval to calibrated, efficient reasoning.

Phase-Interface Instance Segmentation as a Visual Sensor for Laboratory Process Monitoring

Reliable visual monitoring of chemical experiments remains challenging in transparent glassware, where weak phase boundaries and optical artifacts degrade conventional segmentation. We formulate laboratory phenomena as the time evolution of phase interfaces and introduce the Chemical Transparent Glasses dataset 2.0 (CTG 2.0), a vessel-aware benchmark with 3,668 images, 23 glassware categories, and five multiphase interface types for phase-interface instance segmentation. Building on YOLO11m-seg, we propose LGA-RCM-YOLO, which combines Local-Global Attention (LGA) for robust semantic representation and a Rectangular Self-Calibration Module (RCM) for boundary refinement of thin, elongated interfaces. On CTG 2.0, the proposed model achieves 84.4% AP@0.5 and 58.43% AP@0.5-0.95, improving over the YOLO11m baseline by 6.42 and 8.75 AP points, respectively, while maintaining near real-time inference (13.67 FPS, RTX 3060). An auxiliary color-attribute head further labels liquid instances as colored or colorless with 98.71% precision and 98.32% recall. Finally, we demonstrate continuous process monitoring in separatory-funnel phase separation and crystallization, showing that phase-interface instance segmentation can serve as a practical visual sensor for laboratory automation.

Logos: An evolvable reasoning engine for rational molecular design

The discovery and design of functional molecules remain central challenges across chemistry,biology, and materials science. While recent advances in machine learning have accelerated molecular property prediction and candidate generation, existing models tend to excel either in physical fidelity without transparent reasoning, or in flexible reasoning without guarantees of chemical validity. This imbalance limits the reliability of artificial intelligence systems in real scientific design workflows.Here we present Logos, a compact molecular reasoning model that integrates multi-step logical reasoning with strict chemical consistency. Logos is trained using a staged strategy that first exposes the model to explicit reasoning examples linking molecular descriptions to structural decisions, and then progressively aligns these reasoning patterns with molecular representations. In a final training phase, chemical rules and invariants are incorporated directly into the optimization objective, guiding the model toward chemically valid outputs. Across multiple benchmark datasets, Logos achieves strong performance in both structural accuracy and chemical validity, matching or surpassing substantially larger general-purpose language models while operating with a fraction of their parameters. Beyond benchmark evaluation, the model exhibits stable behaviour in molecular optimization tasks involving multiple, potentially conflicting constraints. By explicitly exposing intermediate reasoning steps, Logos enables human inspection and assessment of the design logic underlying each generated structure. These results indicate that jointly optimizing for reasoning structure and physical consistency offers a practical pathway toward reliable and interpretable AI systems for molecular science, supporting closer integration of artificial intelligence into scientific discovery processes.

C$^2$FG: Control Classifier-Free Guidance via Score Discrepancy Analysis

Classifier-Free Guidance (CFG) is a cornerstone of modern conditional diffusion models, yet its reliance on the fixed or heuristic dynamic guidance weight is predominantly empirical and overlooks the inherent dynamics of the diffusion process. In this paper, we provide a rigorous theoretical analysis of the Classifier-Free Guidance. Specifically, we establish strict upper bounds on the score discrepancy between conditional and unconditional distributions at different timesteps based on the diffusion process. This finding explains the limitations of fixed-weight strategies and establishes a principled foundation for time-dependent guidance. Motivated by this insight, we introduce \textbf{Control Classifier-Free Guidance (C$^2$FG)}, a novel, training-free, and plug-in method that aligns the guidance strength with the diffusion dynamics via an exponential decay control function. Extensive experiments demonstrate that C$^2$FG is effective and broadly applicable across diverse generative tasks, while also exhibiting orthogonality to existing strategies.

Preserving Continuous Symmetry in Discrete Spaces: Geometric-Aware Quantization for SO(3)-Equivariant GNNs

Equivariant Graph Neural Networks (GNNs) are essential for physically consistent molecular simulations but suffer from high computational costs and memory bottlenecks, especially with high-order representations. While low-bit quantization offers a solution, applying it naively to rotation-sensitive features destroys the SO(3)-equivariant structure, leading to significant errors and violations of conservation laws. To address this issue, in this work, we propose a Geometric-Aware Quantization (GAQ) framework that compresses and accelerates equivariant models while rigorously preserving continuous symmetry in discrete spaces. Our approach introduces three key contributions: (1) a Magnitude-Direction Decoupled Quantization (MDDQ) scheme that separates invariant lengths from equivariant orientations to maintain geometric fidelity; (2) a symmetry-aware training strategy that treats scalar and vector features with distinct quantization schedules; and (3) a robust attention normalization mechanism to stabilize gradients in low-bit regimes. Experiments on the rMD17 benchmark demonstrate that our W4A8 models match the accuracy of FP32 baselines (9.31 meV vs. 23.20 meV) while reducing Local Equivariance Error (LEE) by over 30x compared to naive quantization. On consumer hardware, GAQ achieves 2.39x inference speedup and 4x memory reduction, enabling stable, energy-conserving molecular dynamics simulations for nanosecond timescales.