Searcharxiv⌕ Search

arXiv subjects

Sihang Liu

Publications and source records attributed to Sihang Liu.

At least 19 recordsLinked to original sources

TileCache: Serving Multi-Round Diffusion Image Editing with Compressed Activation Caches

Multi-round diffusion image editing can reduce computation by reusing intermediate activations, but retaining these activations across editing rounds introduces substantial memory and bandwidth overheads. Compression can reduce these overheads, but efficient cache construction is challenging: compressing only after editing requires buffering over 100 GB of activations per request. Moreover, the compressed cache must support selective reuse, as future masked edits require only a subset of the cached activations. It also exposes a scheduling challenge, as cached edits are faster than initial generations, causing naive policies to severely delay either initial generations or editing requests. In this paper, we aim to address both challenges through compression-informed serving. We develop TileCache, a serving system for multi-round diffusion editing. It compresses reusable activations and incorporates a scheduler that balances both edits and initial generation requests. Compared with the state-of-the-art image editing baseline, TileCache sustains over 2x the request arrival rate at 90% SLO attainment. It completes edits 2.4-4.5x and initial generations 1.3-2.2x faster than the baseline, while preserving image quality.

cs.DC↗

SAKURA: Serving Attention Key-Value State Under Rate Adaptation for Private Image Editing on Phones

Diffusion transformers are widely used in image-editing services. To preserve both efficiency and privacy, existing systems commonly rely on cloud-device collaboration. Keeping request-specific inference on the device avoids sending personal images and prompts to the cloud. These workloads often use common templates, such as in face swapping and virtual try-on, enabling KV state to be precomputed and reused across requests. However, this reusable state is large and scales with the number of templates, making it impractical to store all of them on mobile devices. Storing them in the cloud addresses this capacity issue, but cache transfer is placed onto the inference path and can even be slower than on-device computation, especially under fluctuating mobile networks. We introduce SAKURA, a cloud-edge architecture that delivers reusable attention state to mobile devices under a bandwidth budget. It first uses a fine-grained token-importance-aware compression method to preserve image quality under limited transfer budgets. At runtime, SAKURA adaptively selects the cache representation for each upcoming shard based on network bandwidth. To enable efficient execution, we develop a pipelined runtime that overlaps cache transfer and reconstruction with on-device execution. We evaluate SAKURA using three editing tasks and two diffusion models, using a Snapdragon 8 Elite Gen 5 phone. SAKURA reduces latency by 1.82x and 3.27x over CacheGen adapted to DiT and on-device generation, respectively. It also reduces cache storage by 3.06x over uncompressed cache, while preserving image quality.

cs.SE↗

HybridStitch: Token-Adaptive Model Stitching for Efficient Diffusion Generation

Iterative diffusion generation requires different amounts of model capacity at different token positions. Prior model combinations allocate capacity by denoising step or by drafting and verification role. We present HybridStitch, a training-free framework that uses token-level feedback to select the model supplying each adopted update. Continuous image refinement and discrete language commitment operate on a shared generation state, while model-specific KV buffers provide attention context for selected queries. HybridStitch achieves 1.63--1.68$\times$ faster image generation and 1.15--1.58$\times$ diffusion large language model (dLLM) throughput on four main benchmarks. Estimated dLLM decode query cost decreases by 50--66% with comparable generation quality.

cs.CV↗

Agent-UCT: Upper Confidence Bounds Applied to Trees for Cost-Aware Agentic Workflow Optimization

Workflow search pays for complete quality observations, whose execution cost changes as reusable prefixes accumulate. AgentUCT uses an economic model of the remaining evaluation cost to guide UCT selection while backing up raw quality. Query-prefix accounting connects the search decisions to reusable execution state. For finite fixed-reward trees with bounded adaptive cold-cost estimates and successful persistent caching, we prove an $O(\log T)$ bound on cumulative quality regret over $T$ complete evaluations. Across five paired seeds, AgentUCT achieves 69.57% test F1 versus 68.01% for UCT at a 32M weighted-token budget on a retrospective HotpotQA panel. On GSM8K, it reaches 93.07% mean accuracy at the first budget, compared with 92.60% for UCT and 91.60% for AFlow. Single-factor RAG controls show that online prices and historical continuations improve the selected workflows at equal retained evaluation counts.

cs.AI↗

SparSP: Exploiting Communication Sparsity for Sequence-Parallel Video DiTs

Diffusion Transformers have become the dominant architecture for video generation. Their substantial computational cost motivates scaling inference across multi-GPU servers, yet efficient scaling remains challenging on commodity GPUs connected via PCIe, whose bandwidth is limited. Although sparse attention substantially reduces computation, its implications for communication remain underexplored. This paper argues that attention sparsity should be treated as a communication primitive. We present SparSP, an efficient sparse sequence parallel communication system that co-designs token distribution, communication routing, and asynchronous execution for sparse video diffusion models. First, Dependency-Aware Placement distributes sequence blocks according to diffusion models' sparse attention patterns. Second, Demand-Directed KV Routing transfers KV blocks directly to requesting GPUs without intermediate relays. Third, a Decoupled Transfer Runtime separates communication from GPU computation to reduce resource contention and maximize effective bandwidth. Our evaluation shows that SparSP improves attention performance by 1.38- 1.5$\times$, achieves an average 1.17$\times$ (up to 1.69$\times$) end-to-end speedup across three representative servers and three video diffusion models, and reduces communication volume by 12.54-23.05%. Moreover, we achieve an average 1.53-1.76$\times$ bandwidth improvement over NCCL.

cs.DC↗

AgentDecarbonizer: Carbon-Aware Execution for AI Agents

AI agents extend large language models from single prompt-response interactions to long-running, goaldirected workflows that issue many model calls, invoke tools, and interact with external environments. These workflows enable tasks such as software repair, data analysis, and experiment management, but their repeated model invocations can incur substantial carbon emissions. This paper characterizes the carbon emissions of OpenClaw agent workloads using WildClawBench, and shows that emissions depend on token consumption, context cache reuse, and the carbon intensity of the grid. Our characterization identifies deadline flexibility as an opportunity for carbon-aware execution: agent tasks can wait for lower-carbon-intensity periods or shift to lower-carbon grids. However, doing so requires handling uncertain execution time for temporal shifting and cached context recomputation during spatial shifting. We present AgentDecarbonizer, a carbon optimizer for AI agents that runs alongside OpenClaw. Given a task prompt and user-specified deadline, AgentDecarbonizer conservatively estimates task duration and selects deadline-feasible execution schedules, while accounting for cache recomputation overhead during spatial shifting. Evaluated on WildClawBench workloads with 60 agent tasks across four grids, AgentDecarbonizer reduces carbon emissions by up to 57.9 % compared with a carbon-agnostic baseline and by up to 37.5 % compared with a baseline that selects the carbon-optimal grid at task start time.

cs.LG↗

Active Passivation Tunes Hotspot Locations in GaN Transistors with In Situ Thermal Mechanical Visualization

Efficient thermal dissipation has become critical in emerging electronic devices. However, most existing studies have primarily focused on engineering heat dissipation pathways, largely overlooking the intrinsic behavior of the heat source itself. We demonstrate an active passivation technology that proactively tunes hotspot locations in GaN transistors. By adjusting the active passivation layer length, the hotspot is shifted from the gate edge to the drain-side AP edge, establishing a clear one-to-one spatial correlation. In-situ thermal-mechanical visualization via micro-Raman thermography, combined with multi-physics electro-thermal-mechanical simulations, directly captures the spatial redistribution of both temperature and thermal stress profiles. Electrical analysis confirms that this hotspot migration is driven by the spatial shift of the peak electric field and localized Joule heating. This proactive heat-source tuning strategy provides critical design guidelines for power electronics.

physics.app-ph↗

SpecBench: Evaluating Specification-Level Reasoning for Software Engineering LLM Agents

Software engineering (SWE) agents are transitioning from code generation to full software development lifecycle automation. A critical phase in this lifecycle is specification design: transforming initial proposals into carefully considered requirements through expert review. Existing benchmarks such as SWE-Bench are implementation-focused by measuring the agent's ability to generate code given fixed, precise design requirements. This formulation assumes specifications are correct and complete. In real-world complex and critical software systems, initial specifications are often incomplete and flawed, requiring extensive expert reviews and revisions before being accepted for implementation. To fill this gap, we introduce SpecBench to evaluate specification-level reasoning: the ability to generate complete, unambiguous, consistent, and correct system specifications. SpecBench tasks are derived from the Request for Comments (RFC) process used by mature open-source projects. For each task, an agent is given an initial design proposal, the project codebase, and all past project RFC discussions. The agent is tasked with identifying specification deficiencies: omissions, ambiguities, inconsistencies, or incorrect assumptions in the initial proposal. We evaluate predictions against critiques raised by expert maintainers during historical RFC reviews. SpecBench contains tasks from 5 diverse repositories: Kubernetes, React, Rust, TVM, and vLLM. We evaluate state-of-the-art SWE agents on SpecBench, analyzing their capacity to reason about system design without execution feedback. The best performing agent, GPT-5.4, achieves 44.4% accuracy.

cs.MA↗

Attacks on Approximate Caches in Text-to-Image Diffusion Models

Diffusion models are a powerful class of generative models that produce images and other content from user prompts, but they are computationally intensive. To mitigate this cost, recent academic and industry work has adopted approximate caching, which reuses intermediate states from similar prompts in a cache. While efficient, this optimization introduces new security risks by breaking isolation among users. This paper provides a comprehensive assessment of the security vulnerabilities introduced by approximate caching. First, we demonstrate a remote covert channel established with the approximate cache, where a sender injects prompts with special keywords into the cache system and a receiver can recover that even after days, to exchange information. Second, we introduce a prompt stealing attack using the approximate cache, where an attacker can recover existing cached prompts from hits. Finally, we introduce a poisoning attack that embeds the attacker's logos into the previously stolen prompt, leading to unexpected logo rendering for the requests that hit the poisoned cache prompts. These attacks are all performed remotely through the serving system, demonstrating severe security vulnerabilities in approximate caching. The code for this work is available.

cs.CR↗

Generate "Normal", Edit Poisoned: Branding Injection via Hint Embedding in Image Editing

With the rapid advancement of generative AI, users increasingly rely on image-generation models for image design and creation. To achieve faithful outputs, users typically engage in multi-turn interactions during image refinement: a text-to-image generation phase followed by a text-guided image-to-image editing phase. In this paper, we investigate a novel security vulnerability associated with such a workflow. Our key insight is that a nearly invisible hint, like branding information (e.g., a logo), embedded in an input image can be recognized by downstream generative models and subsequently re-rendered onto semantically related objects, even when the user prompt does not explicitly mention it. This form of hidden payload injection makes the attack stealthy. We study two realistic attack scenarios. The first is a phishing-based setting, in which an attacker controls an online image generation service and injects hidden content into generated images before they are returned to users. The second is a poison-based setting, where an attacker distributes a compromised text-to-image diffusion model whose output contains hidden content. We evaluate both attacks using six injected payloads, including well-known logos and customized designs, and demonstrate that the two attacks can achieve success rates of 44.4% and 32.2% on average, respectively, while ensuring the injected logos are visually imperceptible. We also develop a mitigation solution that achieves an average success rate of 87.4% and 92.3% against the phishing-based and poison-based attacks, respectively.

cs.CR↗

Cache Your Prompt When It's Green: Carbon-Aware Caching for Large Language Model Serving

As large language models (LLMs) become widely used, their environmental impact, especially carbon emission, has attracted more attention. Prior studies focus on compute-related carbon emissions. In this paper, we find that storage is another key contributor. LLM caching, which saves and reuses KV caches for repeated context, reduces operational carbon by avoiding redundant computation. However, this benefit comes at the cost of embodied carbon from high-capacity, high-speed SSDs. As LLMs scale, the embodied carbon of storage grows significantly. To address this tradeoff, we present GreenCache, a carbon-aware cache management framework that dynamically derives resource allocation plans for LLM serving. GreenCache analyzes the correlation between carbon emission and SLO satisfaction, reconfiguring the resource over time to keep the balance between SLO and carbon emission under dynamic workloads. Evaluations from real traces demonstrate that GreenCache achieves an average carbon reduction of 15.1 % when serving Llama-3 70B in the FR grid, with reductions reaching up to 25.3 %, while staying within latency constraints for > 90 % of requests.

cs.DC↗

Side-Channel Attacks on Open vSwitch

Virtualization is widely adopted in cloud systems to manage resource sharing among users. A virtualized environment usually deploys a virtual switch within the host system to enable virtual machines to communicate with each other and with the physical network. The Open vSwitch (OVS) is one of the most popular software-based virtual switches. It maintains a cache hierarchy to accelerate packet forwarding from the host to virtual machines. We characterize the caching system inside OVS from a security perspective and identify three attack primitives. Based on the attack primitives, we present three remote attacks via OVS, breaking the isolation in virtualized environments. First, we identify remote covert channels using different caches. Second, we present a novel header recovery attack that leaks a remote user's packet header fields, breaking the confidentiality guarantees from the system. Third, we demonstrate a remote packet rate monitoring attack that recovers the packet rate of a remote victim. To defend against these attacks, we also discuss and evaluate mitigation solutions.

cs.CR↗

RAID: In-Network RA Signaling Storm Detection for 5G Open RAN

The disaggregation and virtualization of 5G Open RAN (O-RAN) introduces new vulnerabilities in the control plane that can greatly impact the quality of service (QoS) of latency-sensitive 5G applications and services. One critical issue is Random Access (RA) signaling storms where, a burst of illegitimate or misbehaving user equipments (UEs) send Radio Resource Control (RRC) connection requests that rapidly saturate a Central Unit's (CU) processing pipeline. Such storms trigger widespread connection failures within the short contention resolution window defined by 3GPP. Existing detection and mitigation approaches based on near-real-time RAN Intelligent Controller (n-RT RIC) applications cannot guarantee a timely reaction to such attacks as RIC control loops incur tens to hundreds of milliseconds of latency due to the non-deterministic nature of their general purpose processor (GPP) based architectures. This paper presents RAID, an in-network RA signaling storm detection and mitigation system that leverages P4-programmable switch ASICs to enable real-time protection from malicious attacks. RAID embeds a lightweight Random Forest (RF) classifier into a programmable Tofino switch, enabling line-rate flow classification with deterministic microsecond-scale inference delay. By performing ML-based detection directly in the data plane, RAID catches and filters malicious RA requests before they reach and overwhelm the RRC. RAID achieves above 94% detection accuracy with a fixed per-flow inference delay on the order of 3.4 microseconds, effectively meeting strict O-RAN control-plane deadlines. These improvements are sustained across multiple traffic loads, making RAID a fast and scalable solution for the detection and mitigation of signaling storms in 5G O-RAN.

cs.NI↗

A Knowledge Noise Mitigation Framework for Knowledge-based Visual Question Answering

Knowledge-based visual question answering (KB-VQA) requires a model to understand images and utilize external knowledge to provide accurate answers. Existing approaches often directly augment models with retrieved information from knowledge sources while ignoring substantial knowledge redundancy, which introduces noise into the answering process. To address this, we propose a training-free framework with knowledge focusing for KB-VQA, that mitigates the impact of noise by enhancing knowledge relevance and reducing redundancy. First, for knowledge retrieval, our framework concludes essential parts from the image-question pairs, creating low-noise queries that enhance the retrieval of highly relevant knowledge. Considering that redundancy still persists in the retrieved knowledge, we then prompt large models to identify and extract answer-beneficial segments from knowledge. In addition, we introduce a selective knowledge integration strategy, allowing the model to incorporate knowledge only when it lacks confidence in answering the question, thereby mitigating the influence of redundant information. Our framework enables the acquisition of accurate and critical knowledge, and extensive experiments demonstrate that it outperforms state-of-the-art methods.

cs.CV↗

Aegis: Taxonomy and Optimizations for Overcoming Agent-Environment Failures in LLM Agents

Large Language Models (LLMs) agents augmented with domain tools promise to autonomously execute complex tasks requiring human-level intelligence, such as customer service and digital assistance. However, their practical deployment is often limited by their low success rates under complex real-world environments. To tackle this, prior research has primarily focused on improving the agents themselves, such as developing strong agentic LLMs, while overlooking the role of the system environment in which the agent operates. In this paper, we study a complementary direction: improving agent success rates by optimizing the system environment in which the agent operates. We collect 142 agent traces (3,656 turns of agent-environment interactions) across 5 state-of-the-art agentic benchmarks. By analyzing these agent failures, we propose a taxonomy for agent-environment interaction failures that includes 6 failure modes. Guided by these findings, we design Aegis, a set of targeted environment optimizations: 1) environment observability enhancement, 2) common computation offloading, and 3) speculative agentic actions. These techniques improve agent success rates on average by 6.7-12.5%, without any modifications to the agent and underlying LLM.

cs.MA↗

HybridTier: an Adaptive and Lightweight CXL-Memory Tiering System

Modern workloads are demanding increasingly larger memory capacity. Compute Express Link (CXL)-based memory tiering has emerged as a promising solution for addressing this problem by utilizing traditional DRAM alongside slow-tier CXL memory devices. We analyze prior tiering systems and observe two challenges for high-performance memory tiering: adapting to skewed but dynamically varying data hotness distributions while minimizing memory and cache overhead due to tiering. To address these challenges, we propose HybridTier, an adaptive and lightweight tiering system for CXL memory. HybridTier tracks both long-term data access frequency and short-term access momentum \emph{simultaneously} to accurately capture and adapt to shifting hotness distributions. HybridTier reduces the metadata memory overhead by tracking data accesses \emph{probabilistically}, obtaining higher memory efficiency by trading off a small amount of tracking inaccuracy that has a negligible impact on application performance. To reduce cache overhead, HybridTier uses lightweight data structures that optimize for data locality to track data hotness. Our evaluations show that HybridTier outperforms prior systems by up to $91\%$ ($19\%$ geomean), incurring $2.0-7.8\times$ less memory overhead and $1.7-3.5\times$ less cache misses.

cs.DC↗

EnsembleCI: Ensemble Learning for Carbon Intensity Forecasting

Carbon intensity (CI) measures the average carbon emissions generated per unit of electricity, making it a crucial metric for quantifying and managing the environmental impact. Accurate CI predictions are vital for minimizing carbon footprints, yet the state-of-the-art method (CarbonCast) falls short due to its inability to address regional variability and lack of adaptability. To address these limitations, we introduce EnsembleCI, an adaptive, end-to-end ensemble learning-based approach for CI forecasting. EnsembleCI combines weighted predictions from multiple sublearners, offering enhanced flexibility and regional adaptability. In evaluations across 11 regional grids, EnsembleCI consistently surpasses CarbonCast, achieving the lowest mean absolute percentage error (MAPE) in almost all grids and improving prediction accuracy by an average of 19.58%. While performance still varies across grids due to inherent regional diversity, EnsembleCI reduces variability and exhibits greater robustness in long-term forecasting compared to CarbonCast and identifies region-specific key features, underscoring its interpretability and practical relevance. These findings position EnsembleCI as a more accurate and reliable solution for CI forecasting. EnsembleCI source code and data used in this paper are available at https://github.com/emmayly/EnsembleCI.

cs.LG↗

FreeRide: Harvesting Bubbles in Pipeline Parallelism

The occurrence of bubbles in pipeline parallelism is an inherent limitation that can account for more than 40% of the large language model (LLM) training time and is one of the main reasons for the underutilization of GPU resources in LLM training. Harvesting these bubbles for GPU side tasks can increase resource utilization and reduce training costs but comes with challenges. First, because bubbles are discontinuous with various shapes, programming side tasks becomes difficult while requiring excessive engineering effort. Second, a side task can compete with pipeline training for GPU resources and incur significant overhead. To address these challenges, we propose FreeRide, a system designed to harvest bubbles in pipeline parallelism for side tasks. FreeRide provides programmers with interfaces to implement side tasks easily, manages bubbles and side tasks during pipeline training, and controls access to GPU resources by side tasks to reduce overhead. We demonstrate that FreeRide achieves 7.8% average cost savings with a negligible overhead of about 1% in training LLMs while serving model training, graph analytics, and image processing side tasks.

cs.DC↗