The rapid development of large language models (LLMs) has significantly advanced code completion capabilities, giving rise to a new generation of LLM-based Code Completion Tools (LCCTs). Unlike general-purpose LLMs, these tools possess unique workflows, integrating multiple information sources as input and prioritizing code suggestions over natural language interaction, which introduces distinct security challenges. Additionally, LCCTs often rely on proprietary code datasets for training, raising concerns about the potential exposure of sensitive data. This paper exploits these distinct characteristics of LCCTs to develop targeted attack methodologies on two critical security risks: jailbreaking and training data extraction attacks. Our experimental results expose significant vulnerabilities within LCCTs, including a 99.4% success rate in jailbreaking attacks on GitHub Copilot and a 46.3% success rate on Amazon Q. Furthermore, We successfully extracted sensitive user data from GitHub Copilot, including 54 real email addresses and 314 physical addresses associated with GitHub usernames. Our study also demonstrates that these code-based attack methods are effective against general-purpose LLMs, highlighting a broader security misalignment in the handling of code by modern LLMs. These findings underscore critical security challenges associated with LCCTs and suggest essential directions for strengthening their security frameworks.

Eye movements play a significant role in human-computer interaction and are widely recognized as an essential health indicator, making their detection both appealing and technically challenging. In this paper, we present a system named USEE that achieves high-precision capture of weak and aperiodic eye movements by utilizing fine-grained and ubiquitous ultrasound signals, capturing both blinking and more subtle saccades. We first identify signal changes associated with eye movements by capturing the unique impact of blinking. Further, we establish a pioneering relationship between the residuals from signal decomposition and subtle eye movements. Utilizing inno-vative signal processing architectures, we mitigate interference and effectively extract eye movement features. Subsequently, we employ one-dimensional convolutional operations in place of signal cross-correlation, designing filters for motion category identification and a lightweight convolutional neural network for saccade direction classification. This enables our system to serve as a foundational sensing layer for eye movement tracking, applicable across diverse applications. We implement USEE on both a research-purpose platform and a commodity Raspberry Pi. Extensive experimental results demonstrate the effectiveness of our system, achieving 91% accuracy in saccade recognition and 94% in blink detection. The system proves robust, even in challenging scenarios with strong interference, such as the presence of moving pedestrians.

This paper describes our submission to ICASSP 2023 MUG Challenge Track 4, Keyphrase Extraction, which aims to extract keyphrases most relevant to the conference theme from conference materials. We model the challenge as a single-class Named Entity Recognition task and developed techniques for better performance on the challenge: For the data preprocessing, we encode the split keyphrases after word segmentation. In addition, we increase the amount of input information that the model can accept at one time by fusing multiple preprocessed sentences into one segment. We replace the loss function with the multi-class focal loss to address the sparseness of keyphrases. Besides, we score each appearance of keyphrases and add an extra output layer to fit the score to rank keyphrases. Exhaustive evaluations are performed to find the best combination of the word segmentation tool, the pre-trained embedding model, and the corresponding hyperparameters. With these proposals, we scored 45.04 on the final test set.