Qinghua Li

Jiawei Zhang, Xinhao Miao, Jifeng Guo et al.

Planning the motion path for a tightly coupled dual-arm space manipulator under closed-chain constraints is a fundamental yet challenging problem in on-orbit assembly of large-scale space structures. The closed-chain constraints significantly reduce the feasible configuration space, making it difficult for existing planners to efficiently generate collision-free motions, especially in cluttered environments. To address this issue, this paper proposes a task-space constrained bidirectional rapidly-exploring random tree algorithm, termed TCBiRRT. Unlike conventional methods that operate in the high-dimensional configuration space, the proposed approach performs random sampling and node expansion directly in the task space defined by the manipulated object pose. A task-space node expansion strategy is developed to generate candidate object motions, which are then mapped to continuous joint paths using a path inverse kinematics algorithm. The method is further integrated with a bidirectional RRT framework and a regrasp mechanism to efficiently connect two random trees. Extensive simulations are conducted in representative on-orbit assembly scenarios with varying levels of environmental complexity. The results demonstrate that TCBiRRT achieves significantly higher success rates and orders-of-magnitude improvements in planning time compared to state-of-the-art planners. The proposed method provides an efficient and robust solution for motion planning of tightly coupled dual-arm space manipulators.

Jonathan Li, Nasim Farahini, Evgenii Iuliugin et al.

The proliferation of 100B+ parameter Large Language Models (LLMs) with 100k+ context length support have resulted in increasing demands for on-chip memory to support large KV caches. Techniques such as StreamingLLM and SnapKV demonstrate how to control KV cache size while maintaining model accuracy. Yet, these techniques are not commonly used within industrial deployments using frameworks like vLLM or SGLang. The reason is twofold: on one hand, the static graphs and continuous batching methodology employed by these frameworks make it difficult to admit modifications to the standard multi-head attention algorithm, while on the other hand, the accuracy implications of such techniques on modern instruction-following and reasoning models are not well understood, obfuscating the need for implementing these techniques. In this paper, we explore these accuracy implications on Llama-3.1-8B-Instruct and DeepSeek-R1, and develop SnapStream, a KV cache compression method that can be deployed at scale. We demonstrate the efficacy of SnapStream in a 16-way tensor-parallel deployment of DeepSeek-671B on SambaNova SN40L accelerators running at 128k context length and up to 1832 tokens per second in a real production setting. SnapStream enables $4\times$ improved on-chip memory usage and introduces minimal accuracy degradation on LongBench-v2, AIME24 and LiveCodeBench. To the best of our knowledge, this is the first implementation of sparse KV attention techniques deployed in a production inference system with static graphs and continuous batching.

Philip Huff, Dakota Dale, Harshith Guduru et al.

We address a fundamental challenge in cybersecurity operations of translating governance frameworks into actionable mitigation decisions under realistic resource constraints. Frameworks such as the NIST Cybersecurity Framework (CSF) provide widely adopted measures of organizational maturity, but do not directly support the selection and prioritization of defensive strategies against adversarial behavior. We present a system that operationalizes governance frameworks by mapping CSF maturity assessments into MITRE ATT\&CK mitigation capabilities, which enables direct integration of organizational security posture with adversary-informed defensive planning. To manage adversary complexity, we employ a Variable-Order Markov Model (VOMM) trained on observed ATT\&CK technique sequences to enable scalable adversary simulation within a Deep Reinforcement Learning (DRL) environment. We reconstruct likely attack paths and defensive responses using beam search, and then jointly optimize mitigation selection under explicit budget constraints. Our environment supports concurrent adversaries and realistic mitigation costs. Across multiple reward formulations and configurations, we show that the approach produces stable policies, meaningful cost-risk trade-offs, and interpretable mitigation plans aligned with organizational maturity. These results demonstrate that adversary-aware DRL can generate practical, resource-constrained defense strategies grounded in real-world frameworks and threat behavior.

Ang Li, Wei Du, Qinghua Li

Camera is a standard on-board sensor of modern mobile phones. It makes photo taking popular due to its convenience and high resolution. However, when users take a photo of a scenery, a building or a target person, a stranger may also be unintentionally captured in the photo. Such photos expose the location and activity of strangers, and hence may breach their privacy. In this paper, we propose a cooperative mobile photographing scheme called PoliteCamera to protect strangers' privacy. Through the cooperation between a photographer and a stranger, the stranger's face in a photo can be automatically blurred upon his request when the photo is taken. Since multiple strangers nearby the photographer might send out blurring requests but not all of them are in the photo, an adapted balanced convolutional neural network (ABCNN) is proposed to determine whether the requesting stranger is in the photo based on facial attributes. Evaluations demonstrate that the ABCNN can accurately predict facial attributes and PoliteCamera can provide accurate privacy protection for strangers.

Shuhan Yuan, Panpan Zheng, Xintao Wu et al.

Insiders usually cause significant losses to organizations and are hard to detect. Currently, various approaches have been proposed to achieve insider threat detection based on analyzing the audit data that record information of the employee's activity type and time. However, the existing approaches usually focus on modeling the users' activity types but do not consider the activity time information. In this paper, we propose a hierarchical neural temporal point process model by combining the temporal point processes and recurrent neural networks for insider threat detection. Our model is capable of capturing a general nonlinear dependency over the history of all activities by the two-level structure that effectively models activity times, activity types, session durations, and session intervals information. Experimental results on two datasets demonstrate that our model outperforms the models that only consider information of the activity types or time alone.

Wei Du, Ang Li, Qinghua Li

In recent years, machine learning techniques are widely used in numerous applications, such as weather forecast, financial data analysis, spam filtering, and medical prediction. In the meantime, massive data generated from multiple sources further improve the performance of machine learning tools. However, data sharing from multiple sources brings privacy issues for those sources since sensitive information may be leaked in this process. In this paper, we propose a framework enabling multiple parties to collaboratively and accurately train a learning model over distributed datasets while guaranteeing the privacy of data sources. Specifically, we consider logistic regression model for data training and propose two approaches for perturbing the objective function to preserve ε-differential privacy. The proposed solutions are tested on real datasets, including Bank Marketing and Credit Card Default prediction. Experimental results demonstrate that the proposed multiparty learning framework is highly efficient and accurate.

Ang Li, Wei Du, Qinghua Li

The increasing massive data generated by various sources has given birth to big data analytics. Solving large-scale nonlinear programming problems (NLPs) is one important big data analytics task that has applications in many domains such as transport and logistics. However, NLPs are usually too computationally expensive for resource-constrained users. Fortunately, cloud computing provides an alternative and economical service for resource-constrained users to outsource their computation tasks to the cloud. However, one major concern with outsourcing NLPs is the leakage of user's private information contained in NLP formulations and results. Although much work has been done on privacy-preserving outsourcing of computation tasks, little attention has been paid to NLPs. In this paper, we for the first time investigate secure outsourcing of general large-scale NLPs with nonlinear constraints. A secure and efficient transformation scheme at the user side is proposed to protect user's private information; at the cloud side, generalized reduced gradient method is applied to effectively solve the transformed large-scale NLPs. The proposed protocol is implemented on a cloud computing testbed. Experimental evaluations demonstrate that significant time can be saved for users and the proposed mechanism has the potential for practical use.

Ang Li, David Darling, Qinghua Li

Nowadays many people store photos in smartphones. Many of the photos contain sensitive, private information, such as a photocopy of driver's license and credit card. An arising privacy concern is with the unauthorized accesses to such private photos by installed apps. Coarse-grained access control systems such as the Android permission system offer all-or-nothing access to photos stored on smartphones, and users are unaware of the exact behavior of installed apps. Our analysis finds that 82% of the top 200 free apps in a popular Android app store have complete access to stored photos and network on a user's smartphone, which indicates possible private photo leakage. In addition, our user survey reveals that 87.5% of the 112 respondents are not aware that certain apps can access their photos without informing users, and all the respondents believe that the stored photos on their smartphones contain different types of private information. Hence, we propose PhotoSafer, a content-based, context-aware private photo protection system for Android phones. PhotoSafer can detect private photos based on photo content with a well-trained deep convolutional neural network, and control access to photos based on system status (e.g., screen locked or not) and app-running status (e.g., app in the background). Evaluations demonstrate that PhotoSafer can accurately identify private photos in real time. The efficacy and efficiency of the implemented prototype system show the potential for practical use.