Yilin Shen

Shangqian Gao, Ting Hua, Yen-Chang Hsu, Yilin Shen, Hongxia Jin. Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2024.

Shikhar Tuli, Chi-Heng Lin, Yen-Chang Hsu, Niraj Jha, Yilin Shen, Hongxia Jin. Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2024.

Chi-Heng Lin, Shikhar Tuli, James Smith, Yen-Chang Hsu, Yilin Shen, Hongxia Jin. Findings of the Association for Computational Linguistics: NAACL 2024. 2024.

Vision Transformers (ViTs) have emerged as powerful backbones in computer vision, outperforming many traditional CNNs. However, their computational overhead, largely attributed to the self-attention mechanism, makes deployment on resource-constrained edge devices challenging. Multiple solutions rely on token pruning or token merging. In this paper, we introduce "Token Fusion" (ToFu), a method that amalgamates the benefits of both token pruning and token merging. Token pruning proves advantageous when the model exhibits sensitivity to input interpolations, while token merging is effective when the model manifests close to linear responses to inputs. We combine this to propose a new scheme called Token Fusion. Moreover, we tackle the limitations of average merging, which doesn't preserve the intrinsic feature norm, resulting in distributional shifts. To mitigate this, we introduce MLERP merging, a variant of the SLERP technique, tailored to merge multiple tokens while maintaining the norm distribution. ToFu is versatile, applicable to ViTs with or without additional training. Our empirical evaluations indicate that ToFu establishes new benchmarks in both classification and image generation tasks concerning computational efficiency and model accuracy.

Large Language Models (LLMs) are progressively being utilized as machine learning services and interface tools for various applications. However, the security implications of LLMs, particularly in relation to adversarial and Trojan attacks, remain insufficiently examined. In this paper, we propose TrojLLM, an automatic and black-box framework to effectively generate universal and stealthy triggers. When these triggers are incorporated into the input data, the LLMs' outputs can be maliciously manipulated. Moreover, the framework also supports embedding Trojans within discrete prompts, enhancing the overall effectiveness and precision of the triggers' attacks. Specifically, we propose a trigger discovery algorithm for generating universal triggers for various inputs by querying victim LLM-based APIs using few-shot data samples. Furthermore, we introduce a novel progressive Trojan poisoning algorithm designed to generate poisoned prompts that retain efficacy and transferability across a diverse range of models. Our experiments and results demonstrate TrojLLM's capacity to effectively insert Trojans into text prompts in real-world black-box LLM APIs including GPT-3.5 and GPT-4, while maintaining exceptional performance on clean test sets. Our work sheds light on the potential security risks in current models and offers a potential defensive approach. The source code of TrojLLM is available at https://github.com/UCF-ML-Research/TrojLLM.

Keyword spotting systems continuously process audio streams to detect keywords. One of the most challenging tasks in designing such systems is to reduce False Alarm (FA) which happens when the system falsely registers a keyword despite the keyword not being uttered. In this paper, we propose a simple yet elegant solution to this problem that follows from the law of total probability. We show that existing deep keyword spotting mechanisms can be improved by Successive Refinement, where the system first classifies whether the input audio is speech or not, followed by whether the input is keyword-like or not, and finally classifies which keyword was uttered. We show across multiple models with size ranging from 13K parameters to 2.41M parameters, the successive refinement technique reduces FA by up to a factor of 8 on in-domain held-out FA data, and up to a factor of 7 on out-of-domain (OOD) FA data. Further, our proposed approach is "plug-and-play" and can be applied to any deep keyword spotting model.

Water resources in arid and semi-arid inland regions are highly vulnerable, facing threats from global climate change and human activities. Ensuring water resource sustainability requires scientifically evaluating the vulnerability of water resources and its driving factors. Taking the Shule River Basin, an inland river in northwest China, as an example, this study established an assessment system considering 16 influencing factors based on three aspects, including natural vulnerability, anthropogenic vulnerability, and carrying capacity vulnerability. The mature-element fuzzy model based on comprehensive weight and the Delphi method were used to evaluate the water resource vulnerability of the basin from 2005 to 2021. The results were as follows: (1) The water resource vulnerability in the Shule River Basin was between severe and moderate from 2005 to 2021, with a trend towards severe vulnerability. (2) The barrier analysis at the index level shows that factors of natural vulnerability had a low impact on water resource vulnerability in the basin from 2005 to 2019 but exerted a greater impact in 2020–2021, an impact that was far higher than that caused by factors of man-made vulnerability and water resource vulnerability. The impact of factors of anthropogenic vulnerability on water resource vulnerability was relatively low in the entire study period, except in 2016, 2017, 2020, and 2021. In 2005–2010, the factors of bearing capacity vulnerability had a great impact on water resource vulnerability, but in 2011–2021, the impact was alleviated and was gradually reduced. (3) Among the 16 factors affecting water resource vulnerability, the obstacle degree was higher than 6.5% for the following factors: the ratio of irrigation water use on farmland, the annual precipitation, total water resources, the annual quantity of wastewater effluent, the urbanization rate, the surface water control rate, and the degree of groundwater resource amount, exploration, and utilization obstacle values.