Efficiently Serving LLM Reasoning Programs with Certaindex

The rapid evolution of large language models (LLMs) has unlocked their capabilities in advanced reasoning tasks like mathematical problem-solving, code generation, and legal analysis. Central to this progress are inference-time reasoning algorithms, which refine outputs by exploring multiple solution paths, at the cost of increasing compute demands and response latencies. Existing serving systems fail to adapt to the scaling behaviors of these algorithms or the varying difficulty of queries, leading to inefficient resource use and unmet latency targets. We present Dynasor, a system that optimizes inference-time compute for LLM reasoning queries. Unlike traditional engines, Dynasor tracks and schedules requests within reasoning queries and uses Certaindex, a proxy that measures statistical reasoning progress based on model certainty, to guide compute allocation dynamically. Dynasor co-adapts scheduling with reasoning progress: it allocates more compute to hard queries, reduces compute for simpler ones, and terminates unpromising queries early, balancing accuracy, latency, and cost. On diverse datasets and algorithms, Dynasor reduces compute by up to 50% in batch processing and sustaining 3.3x higher query rates or 4.7x tighter latency SLOs in online serving.

Exploiting Hybrid Policy in Reinforcement Learning for Interpretable Temporal Logic Manipulation

Reinforcement Learning (RL) based methods have been increasingly explored for robot learning. However, RL based methods often suffer from low sampling efficiency in the exploration phase, especially for long-horizon manipulation tasks, and generally neglect the semantic information from the task level, resulted in a delayed convergence or even tasks failure. To tackle these challenges, we propose a Temporal-Logic-guided Hybrid policy framework (HyTL) which leverages three-level decision layers to improve the agent's performance. Specifically, the task specifications are encoded via linear temporal logic (LTL) to improve performance and offer interpretability. And a waypoints planning module is designed with the feedback from the LTL-encoded task level as a high-level policy to improve the exploration efficiency. The middle-level policy selects which behavior primitives to execute, and the low-level policy specifies the corresponding parameters to interact with the environment. We evaluate HyTL on four challenging manipulation tasks, which demonstrate its effectiveness and interpretability. Our project is available at: https://sites.google.com/view/hytl-0257/.

Molar: Multimodal LLMs with Collaborative Filtering Alignment for Enhanced Sequential Recommendation

Sequential recommendation (SR) systems have evolved significantly over the past decade, transitioning from traditional collaborative filtering to deep learning approaches and, more recently, to large language models (LLMs). While the adoption of LLMs has driven substantial advancements, these models inherently lack collaborative filtering information, relying primarily on textual content data neglecting other modalities and thus failing to achieve optimal recommendation performance. To address this limitation, we propose Molar, a Multimodal large language sequential recommendation framework that integrates multiple content modalities with ID information to capture collaborative signals effectively. Molar employs an MLLM to generate unified item representations from both textual and non-textual data, facilitating comprehensive multimodal modeling and enriching item embeddings. Additionally, it incorporates collaborative filtering signals through a post-alignment mechanism, which aligns user representations from content-based and ID-based models, ensuring precise personalization and robust performance. By seamlessly combining multimodal content with collaborative filtering insights, Molar captures both user interests and contextual semantics, leading to superior recommendation accuracy. Extensive experiments validate that Molar significantly outperforms traditional and LLM-based baselines, highlighting its strength in utilizing multimodal data and collaborative signals for sequential recommendation tasks. The source code is available at https://anonymous.4open.science/r/Molar-8B06/.

SyNeg: LLM-Driven Synthetic Hard-Negatives for Dense Retrieval

The performance of Dense retrieval (DR) is significantly influenced by the quality of negative sampling. Traditional DR methods primarily depend on naive negative sampling techniques or on mining hard negatives through external retriever and meticulously crafted strategies. However, naive negative sampling often fails to adequately capture the accurate boundaries between positive and negative samples, whereas existing hard negative sampling methods are prone to false negatives, resulting in performance degradation and training instability. Recent advancements in large language models (LLMs) offer an innovative solution to these challenges by generating contextually rich and diverse negative samples. In this work, we present a framework that harnesses LLMs to synthesize high-quality hard negative samples. We first devise a \textit{multi-attribute self-reflection prompting strategy} to direct LLMs in hard negative sample generation. Then, we implement a \textit{hybrid sampling strategy} that integrates these synthetic negatives with traditionally retrieved negatives, thereby stabilizing the training process and improving retrieval performance. Extensive experiments on five benchmark datasets demonstrate the efficacy of our approach, and code is also publicly available.

Fast Causal Discovery by Approximate Kernel-based Generalized Score Functions with Linear Computational Complexity

Score-based causal discovery methods can effectively identify causal relationships by evaluating candidate graphs and selecting the one with the highest score. One popular class of scores is kernel-based generalized score functions, which can adapt to a wide range of scenarios and work well in practice because they circumvent assumptions about causal mechanisms and data distributions. Despite these advantages, kernel-based generalized score functions pose serious computational challenges in time and space, with a time complexity of $\mathcal{O}(n^3)$ and a memory complexity of $\mathcal{O}(n^2)$, where $n$ is the sample size. In this paper, we propose an approximate kernel-based generalized score function with $\mathcal{O}(n)$ time and space complexities by using low-rank technique and designing a set of rules to handle the complex composite matrix operations required to calculate the score, as well as developing sampling algorithms for different data types to benefit the handling of diverse data types efficiently. Our extensive causal discovery experiments on both synthetic and real-world data demonstrate that compared to the state-of-the-art method, our method can not only significantly reduce computational costs, but also achieve comparable accuracy, especially for large datasets.

MotionBridge: Dynamic Video Inbetweening with Flexible Controls

By generating plausible and smooth transitions between two image frames, video inbetweening is an essential tool for video editing and long video synthesis. Traditional works lack the capability to generate complex large motions. While recent video generation techniques are powerful in creating high-quality results, they often lack fine control over the details of intermediate frames, which can lead to results that do not align with the creative mind. We introduce MotionBridge, a unified video inbetweening framework that allows flexible controls, including trajectory strokes, keyframes, masks, guide pixels, and text. However, learning such multi-modal controls in a unified framework is a challenging task. We thus design two generators to extract the control signal faithfully and encode feature through dual-branch embedders to resolve ambiguities. We further introduce a curriculum training strategy to smoothly learn various controls. Extensive qualitative and quantitative experiments have demonstrated that such multi-modal controls enable a more dynamic, customizable, and contextually accurate visual narrative.

Advancing Comprehensive Aesthetic Insight with Multi-Scale Text-Guided Self-Supervised Learning

Image Aesthetic Assessment (IAA) is a vital and intricate task that entails analyzing and assessing an image's aesthetic values, and identifying its highlights and areas for improvement. Traditional methods of IAA often concentrate on a single aesthetic task and suffer from inadequate labeled datasets, thus impairing in-depth aesthetic comprehension. Despite efforts to overcome this challenge through the application of Multi-modal Large Language Models (MLLMs), such models remain underdeveloped for IAA purposes. To address this, we propose a comprehensive aesthetic MLLM capable of nuanced aesthetic insight. Central to our approach is an innovative multi-scale text-guided self-supervised learning technique. This technique features a multi-scale feature alignment module and capitalizes on a wealth of unlabeled data in a self-supervised manner to structurally and functionally enhance aesthetic ability. The empirical evidence indicates that accompanied with extensive instruct-tuning, our model sets new state-of-the-art benchmarks across multiple tasks, including aesthetic scoring, aesthetic commenting, and personalized image aesthetic assessment. Remarkably, it also demonstrates zero-shot learning capabilities in the emerging task of aesthetic suggesting. Furthermore, for personalized image aesthetic assessment, we harness the potential of in-context learning and showcase its inherent advantages.

TrimLLM: Progressive Layer Dropping for Domain-Specific LLMs

Specializing large language models (LLMs) for local deployment in domain-specific use cases is necessary for strong performance while meeting latency and privacy constraints. However, conventional task-specific adaptation approaches do not show simultaneous memory saving and inference speedup at deployment time. Practical compression techniques like quantization and pruning require dedicated hardware or kernel support to achieve measured inference speedup. We develop TrimLLM based on the layer-wise specialization phenomenon we empirically observed and verified on contemporary LLMs. TrimLLM reduces the depth of LLMs via progressive layer dropping. We show it retains LLMs' capacity in specific domains and achieves inference speedup irrespective of hardware and deep learning frameworks. We evaluated TrimLLM on LLMs of various sizes for inference; models adapted on medical, legal, and financial datasets all demonstrate $2.1-5.7\times$ inference speedup on consumer GPUs and up to $3.1\times$ speedup on A100 when compared to state-of-the-art model compression algorithms, with no loss in accuracy at 50$\sim$60\% model compression ratio.

Hesitation and Tolerance in Recommender Systems

User interactions in recommender systems are inherently complex, often involving behaviors that go beyond simple acceptance or rejection. One particularly common behavior is hesitation, where users deliberate over recommended items, signaling uncertainty. Our large-scale surveys, with 6,644 and 3,864 responses respectively, confirm that hesitation is not only widespread but also has a profound impact on user experiences. When users spend additional time engaging with content they are ultimately uninterested in, this can lead to negative emotions, a phenomenon we term as tolerance. The surveys reveal that such tolerance behaviors often arise after hesitation and can erode trust, satisfaction, and long-term loyalty to the platform. For instance, a click might reflect a need for more information rather than genuine interest, and prolonged exposure to unsuitable content amplifies frustration. This misalignment between user intent and system interpretation introduces noise into recommendation training, resulting in suggestions that increase uncertainty and disengagement. To address these issues, we identified signals indicative of tolerance behavior and analyzed datasets from both e-commerce and short-video platforms. The analysis shows a strong correlation between increased tolerance behavior and decreased user activity. We integrated these insights into the training process of a recommender system for a major short-video platform. Results from four independent online A/B experiments demonstrated significant improvements in user retention, achieved with minimal additional computational costs. These findings underscore the importance of recognizing hesitation as a ubiquitous user behavior and addressing tolerance to enhance satisfaction, build trust, and sustain long-term engagement in recommender systems.

Bilevel Learning for Dual-Quadruped Collaborative Transportation under Kinematic and Anisotropic Velocity Constraints

Multi-robot collaborative transportation is a critical capability that has attracted significant attention over recent years. To reliably transport a kinematically constrained payload, a team of robots must closely collaborate and coordinate their individual velocities to achieve the desired payload motion. For quadruped robots, a key challenge is caused by their anisotropic velocity limits, where forward and backward movement is faster and more stable than lateral motion. In order to enable dual-quadruped collaborative transportation and address the above challenges, we propose a novel Bilevel Learning for Collaborative Transportation (BLCT) approach. In the upper-level, BLCT learns a team collaboration policy for the two quadruped robots to move the payload to the goal position, while accounting for the kinematic constraints imposed by their connection to the payload. In the lower-level, BLCT optimizes velocity controls of each individual robot to closely follow the collaboration policy while satisfying the anisotropic velocity constraints and avoiding obstacles. Experiments demonstrate that our BLCT approach well enables collaborative transportation in challenging scenarios and outperforms baseline approaches.