arXiv subjects

Jiasi Chen

Publications and source records attributed to Jiasi Chen.

At least 19 recordsLinked to original sources

arXiv2026

MedCache: Efficient and Temporally Valid Memory for Longitudinal Clinical Agents

Longitudinal clinical agents must maintain an evolving patient state from evidence distributed across visits, time points, and specialties. However, how agent memory should be designed for this setting remains unclear. We introduce a benchmark of multi-visit, multi-specialty patient records that evaluates long-context evidence retrieval, cross-time evidence aggregation, and cross-specialty clinical reasoning. Using this benchmark, we systematically study four memory design choices: curation, organization, retrieval, and memory-augmented reasoning. We find that temporal validity is more important than simply retaining more history; specialty-factorized memory reduces context but can hide shared evidence; and multiple agents help when specialists must reason together, not merely when evidence comes from multiple memories. Guided by these findings, we propose \textit{MedCache}, a hybrid framework that constructs temporally valid patient memory, organizes evidence into overlapping specialty views, routes each query to relevant memories, and adaptively invokes one or multiple specialists. Experiments show that MedCache improves reasoning accuracy and memory efficiency over strong single-agent and multi-agent baselines, while generalizing across model backbones and external datasets.

cs.LG↗

arXiv2026

A Cloud-Edge System for Multimodal Clinical Screening in Resource-Constrained Rural Settings

Medical AI has demonstrated specialist-level diagnostic accuracy, yet these capabilities remain largely inaccessible in resource-constrained rural settings where bandwidth is scarce, compute is limited, and clinical decision-making requires integrating heterogeneous modalities. We introduce a cloud--edge collaborative architecture that addresses these constraints: lightweight, domain-specific models on the edge transform raw medical data into compact structured outputs, while a cloud LLM synthesizes these outputs into clinical summaries. An LLM-based orchestrator dynamically selects diagnostic tools based on patient context, promoting comprehensive modality coverage without processing irrelevant inputs. We evaluate on 20 multimodal clinical cases spanning cardiac, obstetric, trauma, and screening scenarios under three simulated network profiles (500,kbps--5,Mbps). The hybrid system achieves 98--99% diagnostic tool recall with 92--96% precision, matches or exceeds cloud-only baselines on clinical accuracy, and maintains bandwidth-invariant latency (25--35,s) at 4--15x lower token cost. These results highlight the role of architectural design in enabling efficient multimodal integration and improving factual grounding compared to cloud-only approaches under deployment constraints.

cs.LG↗

arXiv2026

GPU-to-Grid: Voltage Regulation via GPU Utilization Control

While the rapid expansion of data centers poses challenges for power grids, it also offers new opportunities as flexible loads. Existing power system research often abstracts data centers as aggregate resources, while computer system research focuses on GPU energy efficiency and largely ignores grid impacts. To bridge this gap, we develop a GPU-to-Grid framework that couples device-level GPU control with power system objectives. We study distribution-level voltage regulation enabled by LLM inference flexibility, using batch size as a data-center-side control knob that trades off GPU power consumption, inference latency, and token throughput. We first formulate the problem as an optimization problem and then realize it as an online feedback optimization controller, implemented by the data center operator using its own empirical GPU power-performance model and real-time measurements from both the GPU and grid systems. Our key insight is that reducing GPU power alleviates lower-voltage violations, while increasing GPU power mitigates upper-voltage violations; this challenges the common belief that minimizing GPU power is always beneficial to power grids.

eess.SY↗

arXiv2026

APE: Agentic Prompt Enhancer for Image Generation and Editing

Natural language has become a powerful interface for image generation and editing, yet text-guided visual systems remain highly sensitive to prompt formulation. Semantically similar requests can produce different outputs depending on wording, specificity, and how explicitly visual constraints are stated, motivating prompt enhancement as a trainable component rather than a peripheral user choice. Existing strong enhancers often rely on large, proprietary LLMs such as ChatGPT or Gemini, adding cost, latency, and deployment dependence to the visual generation pipeline. We propose Agentic Prompt Enhancer (APE), a lightweight framework that post-trains small language models (SLMs) as prompt-enhancement agents. APE supports both single-agent rewriting and role-specialized multi-agent enhancement. Its single-agent instantiation, SAPE, rewrites the prompt in one pass, while its multi-agent instantiation, MAPE, decomposes enhancement into a router--rewriter--composer process for handling compositional constraints over objects, attributes, spatial relations, and edits. With task-aware rewards and post-training protocols, APE improves visual alignment and prompt following without modifying the downstream visual model. Experiments on challenging image generation and editing benchmarks demonstrate that post-trained small prompt enhancers reliably outperform their base counterparts, narrowing the gap to closed-source prompt enhancers; in addition, MAPE proves particularly strong on complex compositional tasks within these benchmarks.

cs.CV↗

arXiv2026

Beyond Trajectory Rewards: Step-level Credit Assignment for Agentic Search via Graph Modeling

In Agentic Search, trajectory-level outcome rewards fail to quantify the behavioral contributions of individual steps, while existing step-level reward methods typically rely on costly tree sampling. We view world knowledge as a latent world graph and each IS task as search within a latent task graph, where effective steps should make graph progress toward the answer node. Based on this prior, we propose Graph-Distance Contribution Reward (GDCR), a step-level process reward that scores newly-retrieved and newly-cited entities by their distance to the answer node in a training-time Entity-Relation (ER) graph. We further propose Step Advantage Policy Optimization (SAPO), which converts GDCR into step-level advantages and combines them with trajectory-level outcome advantages. Experiments on four challenging benchmarks validate the effectiveness of our method.

cs.AI↗

arXiv2026

VSPO: Vector-Steered Policy Optimization for Behavioral Control

Modern language models often need to optimize a primary accuracy objective while also accommodating secondary behavioral preferences, such as verbosity, agreeableness, or the level of technical expertise in its response. In practice, a base model may exhibit a desired behavior very rarely or not at all. Thus, endowing the model with a target behavior creates a sparse behavioral reward bottleneck. To address such multi-objective problems, we introduce Vector-Steered Policy Optimization (VSPO) which employs a steering vector associated with the target behavior to control the behavior intensity of the generated rollouts. VSPO is obtained by modifying GRPO to sample rollouts with varying steering intensities. This process can be interpreted as an on-policy latent self-distillation procedure where the model internalizes its steering vector. By varying steering intensities, VSPO upsamples rare behaviors and enriches rollout diversity, which alleviates the sparse reward issue and provably accelerates the policy optimization. Through comprehensive theory and experiments, we establish that VSPO has favorable properties compared to vanilla reward shaping and other alternative approaches. Specifically, under a bandit abstraction, VSPO provably achieves better iteration complexity than reward-shaped GRPO when the steering-induced distributions are sufficiently aligned with the target behavior. We evaluate VSPO across multiple reasoning benchmarks, including MATH and MMLU-Pro, for four target behaviors: explanation expertise, confidence expression, robustness to misleading context, and response verbosity. Our results show that VSPO consistently improves the control along target behavior while maintaining or improving task accuracy compared with reward shaping, teacher-trace distillation, and guidance-based baselines.

cs.LG↗

arXiv2026

OpenG2G: A Simulation Platform for AI Datacenter-Grid Runtime Coordination

AI's growing compute demand and new datacenter buildouts present major capacity and reliability challenges for the electricity grid, leading to multi-year interconnection delays for new datacenters and bottlenecking AI growth. To ease this strain, datacenters increasingly offer rapid power flexibility in response to grid signals, where the datacenter can increase or decrease its power consumption by adapting its workload in real time. In order to understand the impact of large datacenters on the grid and to facilitate the design of effective coordination strategies, we build OpenG2G, a simulation platform for AI datacenter-grid runtime coordination. We show that OpenG2G is capable of answering a wide range of coordination questions by allowing users to implement and compare various control paradigms (including classic, optimization, and learning-based controllers), and quantify how AI model and deployment choices affect datacenter flexibility and coordination outcomes. This versatility is enabled by OpenG2G's modular and extensible architecture: a datacenter backend driven by real measurements of production-grade AI services, a grid backend built on high-fidelity grid simulators, and a generic controller interface that closes the loop between them. We describe the design of OpenG2G and demonstrate its usefulness through realistic grid scenarios and AI workloads.

cs.LG↗

arXiv2025

SmartChunk Retrieval: Query-Aware Chunk Compression with Planning for Efficient Document RAG

Retrieval-augmented generation (RAG) has strong potential for producing accurate and factual outputs by combining language models (LMs) with evidence retrieved from large text corpora. However, current pipelines are limited by static chunking and flat retrieval: documents are split into short, predetermined, fixed-size chunks, embeddings are retrieved uniformly, and generation relies on whatever chunks are returned. This design brings challenges, as retrieval quality is highly sensitive to chunk size, often introduces noise from irrelevant or misleading chunks, and scales poorly to large corpora. We present SmartChunk retrieval, a query-adaptive framework for efficient and robust long-document question answering (QA). SmartChunk uses (i) a planner that predicts the optimal chunk abstraction level for each query, and (ii) a lightweight compression module that produces high-level chunk embeddings without repeated summarization. By adapting retrieval granularity on the fly, SmartChunk balances accuracy with efficiency and avoids the drawbacks of fixed strategies. Notably, our planner can reason about chunk abstractions through a novel reinforcement learning scheme, STITCH, which boosts accuracy and generalization. To reflect real-world applications, where users face diverse document types and query styles, we evaluate SmartChunk on five QA benchmarks plus one out-of-domain dataset. Across these evaluations, SmartChunk outperforms state-of-the-art RAG baselines, while reducing cost. Further analysis demonstrates strong scalability with larger corpora and consistent gains on out-of-domain datasets, highlighting its effectiveness as a general framework for adaptive retrieval.

cs.IR↗

arXiv2025

Evil Vizier: Vulnerabilities of LLM-Integrated XR Systems

Extended reality (XR) applications increasingly integrate Large Language Models (LLMs) to enhance user experience, scene understanding, and even generate executable XR content, and are often called "AI glasses". Despite these potential benefits, the integrated XR-LLM pipeline makes XR applications vulnerable to new forms of attacks. In this paper, we analyze LLM-Integated XR systems in the literature and in practice and categorize them along different dimensions from a systems perspective. Building on this categorization, we identify a common threat model and demonstrate a series of proof-of-concept attacks on multiple XR platforms that employ various LLM models (Meta Quest 3, Meta Ray-Ban, Android, and Microsoft HoloLens 2 running Llama and GPT models). Although these platforms each implement LLM integration differently, they share vulnerabilities where an attacker can modify the public context surrounding a legitimate LLM query, resulting in erroneous visual or auditory feedback to users, thus compromising their safety or privacy, sowing confusion, or other harmful effects. To defend against these threats, we discuss mitigation strategies and best practices for developers, including an initial defense prototype, and call on the community to develop new protection mechanisms to mitigate these risks.

cs.CR↗

arXiv2025

BREAD: Branched Rollouts from Expert Anchors Bridge SFT & RL for Reasoning

Small language models (SLMs) struggle to learn complex reasoning behaviors, especially when high-quality traces are scarce or difficult to learn from. The standard training approach combines a supervised fine-tuning (SFT) stage, often to distill capabilities of a larger model, followed by a reinforcement learning (RL)stage such as Group Relative Policy Optimization (GRPO). In this paper, we investigate the fundamental limitations of this SFT + RL paradigm and propose methods to overcome them. Under a suitable theoretical model, we demonstrate that the SFT + RL strategy can fail completely when (1) the expert's traces are too difficult for the small model to express, or (2) the small model's initialization has exponentially small likelihood of success. To address these, we introduce BREAD: a GRPO variant that unifies the SFT and RL stages via partial expert guidance and branched rollouts. When self-generated traces fail, BREAD adaptively inserts short expert prefixes/hints, allowing the small model to complete the rest of the reasoning path, and ensuring that each update includes at least one successful trace. This mechanism both densifies the reward signal and induces a natural learning curriculum. BREAD requires fewer than 40% of ground-truth traces, consistently outperforming standard GRPO while speeding up the training by about 3 times. Importantly, we demonstrate that BREAD helps the model solve problems that are otherwise unsolvable by the SFT + RL strategy, highlighting how branched rollouts and expert guidance can substantially boost SLM reasoning.

cs.LG↗

arXiv2025

Making Small Language Models Efficient Reasoners: Intervention, Supervision, Reinforcement

Recent research enhances language model reasoning by scaling test-time compute via longer chain-of-thought traces. This often improves accuracy but also introduces redundancy and high computational cost, especially for small language models distilled with supervised fine-tuning (SFT). In this work, we propose new algorithms to improve token-efficient reasoning with small-scale models by effectively trading off accuracy and computation. We first show that the post-SFT model fails to determine the optimal stopping point of the reasoning process, resulting in verbose and repetitive outputs. Verbosity also significantly varies across wrong vs correct responses. To address these issues, we propose two solutions: (1) Temperature scaling (TS) to control the stopping point for the thinking phase and thereby trace length, and (2) TLDR: a length-regularized reinforcement learning method based on GRPO that facilitates multi-level trace length control (e.g. short, medium, long reasoning). Experiments on four reasoning benchmarks, MATH500, AMC, AIME24 and OlympiadBench, demonstrate that TS is highly effective compared to s1's budget forcing approach and TLDR significantly improves token efficiency by about 50% with minimal to no accuracy loss over the SFT baseline. Moreover, TLDR also facilitates flexible control over the response length, offering a practical and effective solution for token-efficient reasoning in small models. Ultimately, our work reveals the importance of stopping time control, highlights shortcomings of pure SFT, and provides effective algorithmic recipes.

cs.LG↗

arXiv2025

L3GS: Layered 3D Gaussian Splats for Efficient 3D Scene Delivery

Traditional 3D content representations include dense point clouds that consume large amounts of data and hence network bandwidth, while newer representations such as neural radiance fields suffer from poor frame rates due to their non-standard volumetric rendering pipeline. 3D Gaussian splats (3DGS) can be seen as a generalization of point clouds that meet the best of both worlds, with high visual quality and efficient rendering for real-time frame rates. However, delivering 3DGS scenes from a hosting server to client devices is still challenging due to high network data consumption (e.g., 1.5 GB for a single scene). The goal of this work is to create an efficient 3D content delivery framework that allows users to view high quality 3D scenes with 3DGS as the underlying data representation. The main contributions of the paper are: (1) Creating new layered 3DGS scenes for efficient delivery, (2) Scheduling algorithms to choose what splats to download at what time, and (3) Trace-driven experiments from users wearing virtual reality headsets to evaluate the visual quality and latency. Our system for Layered 3D Gaussian Splats delivery L3GS demonstrates high visual quality, achieving 16.9% higher average SSIM compared to baselines, and also works with other compressed 3DGS representations.

cs.GR↗

arXiv2025

Leveraging Audio Representations for Vibration-Based Crowd Monitoring in Stadiums

Crowd monitoring in sports stadiums is important to enhance public safety and improve the audience experience. Existing approaches mainly rely on cameras and microphones, which can cause significant disturbances and often raise privacy concerns. In this paper, we sense floor vibration, which provides a less disruptive and more non-intrusive way of crowd sensing, to predict crowd behavior. However, since the vibration-based crowd monitoring approach is newly developed, one main challenge is the lack of training data due to sports stadiums being large public spaces with complex physical activities. In this paper, we present ViLA (Vibration Leverage Audio), a vibration-based method that reduces the dependency on labeled data by pre-training with unlabeled cross-modality data. ViLA is first pre-trained on audio data in an unsupervised manner and then fine-tuned with a minimal amount of in-domain vibration data. By leveraging publicly available audio datasets, ViLA learns the wave behaviors from audio and then adapts the representation to vibration, reducing the reliance on domain-specific vibration data. Our real-world experiments demonstrate that pre-training the vibration model using publicly available audio data (YouTube8M) achieved up to a 5.8x error reduction compared to the model without audio pre-training.

cs.SD↗

arXiv2025

Siren Song: Manipulating Pose Estimation in XR Headsets Using Acoustic Attacks

Extended Reality (XR) experiences involve interactions between users, the real world, and virtual content. A key step to enable these experiences is the XR headset sensing and estimating the user's pose in order to accurately place and render virtual content in the real world. XR headsets use multiple sensors (e.g., cameras, inertial measurement unit) to perform pose estimation and improve its robustness, but this provides an attack surface for adversaries to interfere with the pose estimation process. In this paper, we create and study the effects of acoustic attacks that create false signals in the inertial measurement unit (IMU) on XR headsets, leading to adverse downstream effects on XR applications. We generate resonant acoustic signals on a HoloLens 2 and measure the resulting perturbations in the IMU readings, and also demonstrate both fine-grained and coarse attacks on the popular ORB-SLAM3 and an open-source XR system (ILLIXR). With the knowledge gleaned from attacking these open-source frameworks, we demonstrate four end-to-end proof-of-concept attacks on a HoloLens 2: manipulating user input, clickjacking, zone invasion, and denial of user interaction. Our experiments show that current commercial XR headsets are susceptible to acoustic attacks, raising concerns for their security.

cs.CR↗

arXiv2024

Efficient Contextual LLM Cascades through Budget-Constrained Policy Learning

Recent successes in natural language processing have led to the proliferation of large language models (LLMs) by multiple providers. Each LLM offering has different inference accuracy, monetary cost, and latency, and their accuracy further depends on the exact wording of the question (i.e., the specific prompt). At the same time, users often have a limit on monetary budget and latency to answer all their questions, and they do not know which LLMs to choose for each question to meet their accuracy and long term budget requirements. To navigate this rich design space, we propose TREACLE ($\underline{T}$hrifty $\underline{Rea}$soning via $\underline{C}$ontext-Aware $\underline{L}$LM and Prompt S$\underline{e}$lection), a reinforcement learning policy that jointly selects the model and prompting scheme while respecting the user's monetary cost and latency constraints. TREACLE uses the problem context, including question text embeddings (reflecting the type or difficulty of a query) and the response history (reflecting the consistency of previous responses) to make smart decisions. Our evaluations on standard reasoning datasets (GSM8K, CSQA, and LLC) with various LLMs and prompts show that TREACLE enables cost savings of up to 85% compared to baselines, while maintaining high accuracy. Importantly, it provides the user with the ability to gracefully trade off accuracy for cost.

cs.CL↗

arXiv2024

Selective Attention: Enhancing Transformer through Principled Context Control

The attention mechanism within the transformer architecture enables the model to weigh and combine tokens based on their relevance to the query. While self-attention has enjoyed major success, it notably treats all queries $q$ in the same way by applying the mapping $V^\top\text{softmax}(Kq)$, where $V,K$ are the value and key embeddings respectively. In this work, we argue that this uniform treatment hinders the ability to control contextual sparsity and relevance. As a solution, we introduce the $\textit{Selective Self-Attention}$ (SSA) layer that augments the softmax nonlinearity with a principled temperature scaling strategy. By controlling temperature, SSA adapts the contextual sparsity of the attention map to the query embedding and its position in the context window. Through theory and experiments, we demonstrate that this alleviates attention dilution, aids the optimization process, and enhances the model's ability to control softmax spikiness of individual queries. We also incorporate temperature scaling for value embeddings and show that it boosts the model's ability to suppress irrelevant/noisy tokens. Notably, SSA is a lightweight method which introduces less than 0.5% new parameters through a weight-sharing strategy and can be fine-tuned on existing LLMs. Extensive empirical evaluations demonstrate that SSA-equipped models achieve a noticeable and consistent accuracy improvement on language modeling benchmarks.

cs.LG↗

arXiv2024

That Doesn't Go There: Attacks on Shared State in Multi-User Augmented Reality Applications

Augmented Reality (AR) is expected to become a pervasive component in enabling shared virtual experiences. In order to facilitate collaboration among multiple users, it is crucial for multi-user AR applications to establish a consensus on the "shared state" of the virtual world and its augmentations, through which they interact within augmented reality spaces. Current methods to create and access shared state collect sensor data from devices (e.g., camera images), process them, and integrate them into the shared state. However, this process introduces new vulnerabilities and opportunities for attacks. Maliciously writing false data to "poison" the shared state is a major concern for the security of the downstream victims that depend on it. Another type of vulnerability arises when reading the shared state; by providing false inputs, an attacker can view hologram augmentations at locations they are not allowed to access. In this work, we demonstrate a series of novel attacks on multiple AR frameworks with shared states, focusing on three publicly-accessible frameworks. We show that these frameworks, while using different underlying implementations, scopes, and mechanisms to read from and write to the shared state, have shared vulnerability to a unified threat model. Our evaluation of these state-of-art AR applications demonstrates reliable attacks both on updating and accessing shared state across the different systems. To defend against such threats, we discuss a number of potential mitigation strategies that can help enhance the security of multi-user AR applications.

cs.CR↗

arXiv2024

Class-attribute Priors: Adapting Optimization to Heterogeneity and Fairness Objective

Modern classification problems exhibit heterogeneities across individual classes: Each class may have unique attributes, such as sample size, label quality, or predictability (easy vs difficult), and variable importance at test-time. Without care, these heterogeneities impede the learning process, most notably, when optimizing fairness objectives. Confirming this, under a gaussian mixture setting, we show that the optimal SVM classifier for balanced accuracy needs to be adaptive to the class attributes. This motivates us to propose CAP: An effective and general method that generates a class-specific learning strategy (e.g. hyperparameter) based on the attributes of that class. This way, optimization process better adapts to heterogeneities. CAP leads to substantial improvements over the naive approach of assigning separate hyperparameters to each class. We instantiate CAP for loss function design and post-hoc logit adjustment, with emphasis on label-imbalanced problems. We show that CAP is competitive with prior art and its flexibility unlocks clear benefits for fairness objectives beyond balanced accuracy. Finally, we evaluate CAP on problems with label noise as well as weighted test objectives to showcase how CAP can jointly adapt to different heterogeneities.

cs.LG↗