Searcharxiv⌕ Search

arXiv subjects

Jinxin Yang

Publications and source records attributed to Jinxin Yang.

7 recordsLinked to original sources

Cumulant dynamics of the ferromagnetic dipolar XY model

We study the second-order truncated cumulant dynamics for the preparation of dynamical states in the dipolar ferromagnetic XY model on the square lattice for system sizes up to linear size L = 60. The observations lead to a transparent picture in which a coherent region of finite extent and finite magnetization forms, which is possible even for diabatic but sufficiently slow ramps. These two scales are functions of the ramp constant, the initial entropy of the system, and time. Our findings put strong requirements on future experiments addressing dipolar XY equilibrium physics in bigger system sizes.

cond-mat.stat-mech↗

SAFE-Mem: Secure Agentic Memory with Fine-Grained Authority and Governance-Aware Retrieval

Large language model (LLM) agents increasingly rely on persistent memory to support long-term reasoning and personalized interactions. However, information collected from external sources can introduce security risks when untrusted claims (i.e., atomic statements) are later treated as authorized preferences, policies, or permissions. Existing defenses typically rely on content-based filtering or coarse-grained provenance tracking, which either fail to prevent unauthorized memory use or block legitimate information, sacrificing memory utility. To address these limitations, we propose SAFE-Mem, a fine-grained memory governance framework that improves security while preserving legitimate memory use. First, SAFE-Mem annotates each claim with three governance labels (i.e., kind, authority, and scope) to capture its provenance and permitted use. Second, governance-aware retrieval matches query requirements with claim permissions to block unauthorized influence while retaining legitimate information. Third, a governance-enabled memory graph connects source sentences, claims, and entities to support provenance tracking and efficient retrieval. Across four LLM backbones on MPBench and LoCoMo, SAFE-Mem attains the lowest attack success among defenses that retain legitimate memory, at least 77% below the strongest of them, and the highest answer accuracy of all evaluated defenses, improving it by up to 15 points. The code is available at https://github.com/LVenum/SAFE-Mem.

cs.CR↗

CUA-Recovery: Benchmarking Error Recovery in Long-Horizon Computer-Use Agents

As computer-use agents (CUAs) move to long-horizon workflows spanning multiple applications, undetected errors can silently corrupt persistent state and cause cascading failures in subsequent actions. Reliable error detection and recovery have become critical bottlenecks for building robust CUAs. However, existing benchmarks focus primarily on end-to-end task performance, providing limited insight into whether agents can detect and recover from errors during execution. The few benchmarks that evaluate recovery rely largely on manually injected errors, which may not reflect the types or horizons of failures induced by real agent policies. We introduce CUA-Recovery, a benchmark comprising test and training suites across multiple desktop applications. We construct it using ReRail, a framework for synthesizing long-horizon tasks and trajectories in a persistent desktop environment. ReRail composes MyPCBench tasks into new long-horizon workflows with executable ground truth, collects verifier-confirmed failures from multiple policies, and turns them into verified recovery trajectories. CUA-Recovery comprises a human-verified test suite of 2,250 erroneous states over 100 tasks and a training suite of 8.6K trajectories over 377 tasks. Evaluation on the test suite shows that agents struggle both to recognize inherited errors and to complete the task. Across six agents and six takeover depths, Error Awareness Rate (EAR) reaches at most 44.0% and Pass@3 at most 25.5%. After fine-tuning Qwen3.5-35B-A3B on the training suite, its average EAR rises from 19.9% to 35.2% on held-out recovery states, and its Rubric Score goes from 5.0% to 16.1%. It also gains 15.3 points on MyPCBench and 10.8 points on Odysseys for Pass@3, showing that the training trajectories help both with recovery and with completing tasks from a clean start. Our code is available at https://github.com/xu-hu-2002/CUA-Recovery.

cs.AI↗

Anatomy of Agentic Memory: Taxonomy and Empirical Analysis of Evaluation and System Limitations

Agentic memory systems enable large language model (LLM) agents to maintain state across long interactions, supporting long-horizon reasoning and personalization beyond fixed context windows. Despite rapid architectural development, the empirical foundations of these systems remain fragile: existing benchmarks are often underscaled, evaluation metrics are misaligned with semantic utility, performance varies significantly across backbone models, and system-level costs are frequently overlooked. This survey presents a structured analysis of agentic memory from both architectural and system perspectives. We first introduce a concise taxonomy of MAG systems based on four memory structures. Then, we analyze key pain points limiting current systems, including benchmark saturation effects, metric validity and judge sensitivity, backbone-dependent accuracy, and the latency and throughput overhead introduced by memory maintenance. By connecting the memory structure to empirical limitations, this survey clarifies why current agentic memory systems often underperform their theoretical promise and outlines directions for more reliable evaluation and scalable system design.

cs.CL↗

Status Updating in Two-Way Delay Systems with Preemption

We consider a status update system consisting of a sampler, a sink, and a controller located at the sink. The controller sends requests to the sampler to generate and transmit status updates. Packet transmissions from the controller to the sampler (reverse link) and from the sampler to the sink (forward link) experience random delays. The reverse and forward links are modeled as servers with geometric service times, referred to as the controller and sampler servers, respectively. Each server is equipped with a single buffer that stores an arriving packet when the server is busy. We adopt a preemption-in-waiting policy on both links, whereby an arriving packet replaces the packet in the buffer whenever the buffer is full. Our main goal is to determine the optimal generation times of request packets at the controller in order to minimize the long-term average age of information (AoI) at the sink. We formulate the problem as a Markov decision process (MDP) and derive the optimal stationary deterministic policy using the relative value iteration (RVI) algorithm. We prove the convergence of the algorithm. Numerical results show that the proposed system consistently outperforms baseline policies from prior work and reveal a threshold-based structure for the optimal policy.

cs.IT↗

Bala-Join: An Adaptive Hash Join for Balancing Communication and Computation in Geo-Distributed SQL Databases

Shared-nothing geo-distributed SQL databases, such as CockroachDB, are increasingly vital for enterprise applications requiring data resilience and locality. However, we encountered significant performance degradation at the customer side, especially when their deployments span multiple data centers over a Wide Area Network (WAN). Our investigation identifies the bottleneck in the performance of the Distributed Hash Join (Dist-HJ) algorithm, which is contingent upon a crucial balance between communication overhead and computational load. This balance is severely disrupted when processing skewed data from real-world customer workloads, leading to the observed performance decline. To tackle this challenge, we introduce Bala-Join, an adaptive solution to balance the computation and network load in Dist-HJ execution. Our approach consists of the Balanced Partition and Partial Replication (BPPR) algorithm and a distributed online skewed join key detector. The former achieves balanced redistribution of skewed data through a multicast mechanism to improve computational performance and reduce network overhead. The latter provides real-time skewed join key information tailored to BPPR. Furthermore, an Active-Signaling and Asynchronous-Pulling (ASAP) mechanism is incorporated to enable efficient, real-time synchronization between the detector and the redistribution process with minimal overhead. Empirical study shows that Bala-Join outperforms the popular Dist-HJ solutions, increasing throughput by 25%-61%.

cs.DB↗

One Size Cannot Fit All: a Self-Adaptive Dispatcher for Skewed Hash Join in Shared-nothing RDBMSs

Shared-nothing architecture has been widely adopted in various commercial distributed RDBMSs. Thanks to the architecture, query can be processed in parallel and accelerated by scaling up the cluster horizontally on demand. In spite of that, load balancing has been a challenging issue in all distributed RDBMSs, including shared-nothing ones, which suffers much from skewed data distribution. In this work, we focus on one of the representative operator, namely Hash Join, and investigate how skewness among the nodes of a cluster will affect the load balance and eventual efficiency of an arbitrary query in shared-nothing RDBMSs. We found that existing Distributed Hash Join (Dist-HJ) solutions may not provide satisfactory performance when a value is skewed in both the probe and build tables. To address that, we propose a novel Dist-HJ solution, namely Partition and Replication (PnR). Although PnR provide the best efficiency in some skewness scenario, our exhaustive experiments over a group of shared-nothing RDBMSs show that there is not a single Dist-HJ solution that wins in all (data skew) scenarios. To this end, we further propose a self-adaptive Dist-HJ solution with a builtin sub-operator cost model that dynamically select the best Dist-HJ implementation strategy at runtime according to the data skew of the target query. We implement the solution in our commercial shared-nothing RDBMSs, namely KaiwuDB (former name ZNBase) and empirical study justifies that the self-adaptive model achieves the best performance comparing to a series of solution adopted in many existing RDBMSs.

cs.DB↗