Search

Your search keyword '"Pan, Zihao"' showing total 430 results
Start Over Author "Pan, Zihao"
430 results on '"Pan, Zihao"'

1. BiasEye: A Bias-Aware Real-time Interactive Material Screening System for Impartial Candidate Assessment

Author

Liu, Qianyu, Jiang, Haoran, Pan, Zihao, Han, Qiushi, Peng, Zhenhui, and Li, Quan

Subjects
Computer Science - Human-Computer Interaction
Abstract

In the process of evaluating competencies for job or student recruitment through material screening, decision-makers can be influenced by inherent cognitive biases, such as the screening order or anchoring information, leading to inconsistent outcomes. To tackle this challenge, we conducted interviews with seven experts to understand their challenges and needs for support in the screening process. Building on their insights, we introduce BiasEye, a bias-aware real-time interactive material screening visualization system. BiasEye enhances awareness of cognitive biases by improving information accessibility and transparency. It also aids users in identifying and mitigating biases through a machine learning (ML) approach that models individual screening preferences. Findings from a mixed-design user study with 20 participants demonstrate that, compared to a baseline system lacking our bias-aware features, BiasEye increases participants' bias awareness and boosts their confidence in making final decisions. At last, we discuss the potential of ML and visualization in mitigating biases during human decision-making tasks., Comment: IUI' 24, March 18-21, 2024, Greenville, SC, USA

Published
2024
Full Text
View/download PDF

2. It Ain't That Bad: Understanding the Mysterious Performance Drop in OOD Generalization for Generative Transformer Models

Author

Xu, Xingcheng, Pan, Zihao, Zhang, Haipeng, and Yang, Yanqing

Subjects
Computer Science - Machine Learning, Computer Science - Artificial Intelligence
Abstract

Large language models (LLMs) have achieved remarkable proficiency on solving diverse problems. However, their generalization ability is not always satisfying and the generalization problem is common for generative transformer models in general. Researchers take basic mathematical tasks like n-digit addition or multiplication as important perspectives for investigating their generalization behaviors. It is observed that when training models on n-digit operations (e.g., additions) in which both input operands are n-digit in length, models generalize successfully on unseen n-digit inputs (in-distribution (ID) generalization), but fail miserably on longer, unseen cases (out-of-distribution (OOD) generalization). We bring this unexplained performance drop into attention and ask whether there is systematic OOD generalization. Towards understanding LLMs, we train various smaller language models which may share the same underlying mechanism. We discover that the strong ID generalization stems from structured representations, while behind the unsatisfying OOD performance, the models still exhibit clear learned algebraic structures. Specifically, these models map unseen OOD inputs to outputs with learned equivalence relations in the ID domain, which we call the equivalence generalization. These findings deepen our knowledge regarding the generalizability of generative models including LLMs, and provide insights into potential avenues for improvement., Comment: Proceedings of the 33rd International Joint Conference on Artificial Intelligence (IJCAI-24)

Published
2023

3. For the Underrepresented in Gender Bias Research: Chinese Name Gender Prediction with Heterogeneous Graph Attention Network

Author

Pan, Zihao, Peng, Kai, Ling, Shuai, and Zhang, Haipeng

Subjects
Computer Science - Artificial Intelligence
Abstract

Achieving gender equality is an important pillar for humankind's sustainable future. Pioneering data-driven gender bias research is based on large-scale public records such as scientific papers, patents, and company registrations, covering female researchers, inventors and entrepreneurs, and so on. Since gender information is often missing in relevant datasets, studies rely on tools to infer genders from names. However, available open-sourced Chinese gender-guessing tools are not yet suitable for scientific purposes, which may be partially responsible for female Chinese being underrepresented in mainstream gender bias research and affect their universality. Specifically, these tools focus on character-level information while overlooking the fact that the combinations of Chinese characters in multi-character names, as well as the components and pronunciations of characters, convey important messages. As a first effort, we design a Chinese Heterogeneous Graph Attention (CHGAT) model to capture the heterogeneity in component relationships and incorporate the pronunciations of characters. Our model largely surpasses current tools and also outperforms the state-of-the-art algorithm. Last but not least, the most popular Chinese name-gender dataset is single-character based with far less female coverage from an unreliable source, naturally hindering relevant studies. We open-source a more balanced multi-character dataset from an official source together with our code, hoping to help future research promoting gender equality., Comment: 8 pages, 4 figures

Published
2023

4. Changes in University Students' Behavioral Intention to Learn Online throughout the COVID-19: Insights for Online Teaching in the Post-Pandemic Era

Author

Zhu, Yue, Geng, Gretchen, Disney, Leigh, and Pan, Zihao

Abstract

Many researchers investigated university students' behavioural intention to undertake online courses during COVID. However, few examined how students' intention might change throughout COVID by incorporating their learning capability and approaches. The universities in China went through a process from lockdown in February to reopening in September 2020. It provided a unique context for university students in China to experience emergent online learning for approximately six months before returning to normal face-to-face or blended learning on campus. The researchers conducted a questionnaire survey among 193 Chinese university students to investigate the changes in their behavioral intention to learn online throughout COVID. Additionally, the researchers explored the relationships between the participants' behavioral intention and the factors of learning capability in general, application of specific online learning strategies, online course engagement levels, and academic performance. It was found that the participants' intention to study online significantly increased during COVID and then slightly decreased after the university reopened. The participants' intention of online learning after COVID was predicted by their prior intention, learning capability, application of online learning strategies, and online course engagement. The participants' perceptions about online learning revealed that, when choosing future course delivery modes, they would a) reflect on their own disposition, capability, and needs, b) compare different learning modes, and c) examine course quality and teachers' competency. The participants also shared advice regarding their expectation of future online courses which may help shape university educators' pedagogical practices and provide insights for university online and blended course delivery from learners' perspectives.

Published
2023
Full Text
View/download PDF

32. Engineered Extracellular Vesicles from LncEEF1G Overexpressing Mesenchymal Stem Cells Promote Fibrotic Liver Regeneration by Upregulating HGF Release of Hepatic Stellate Cells

Author

Zheng, Jun, primary, Zhang, Jiebin, additional, Lei, Yunguo, additional, Liu, Xiaoquan, additional, Qiu, Xiaotong, additional, Chen, Haitian, additional, Li, Rong, additional, Wang, Tingting, additional, Sui, Xin, additional, Xiao, Jiaqi, additional, Jiang, Chenhao, additional, Zhang, Huayao, additional, Liu, Yasong, additional, Zhang, Yingcai, additional, Yao, Jia, additional, Che, Xu, additional, Lin, Ye, additional, and Pan, Zihao, additional

Published
2024
Full Text
View/download PDF

45. Type II and IV toxin-antitoxin systems coordinately stabilize the integrative and conjugative element of the ICESa2603 family conferring multiple drug resistance in Streptococcus suis.

Author

Gu, Qibing, Zhu, Xiayu, Yu, Yong, Jiang, Tao, Pan, Zihao, Ma, Jiale, and Yao, Huochun

Subjects
MULTIDRUG resistance, STREPTOCOCCUS suis, MOBILE genetic elements, PALINDROMIC DNA, DRUG stability
Abstract

Integrative and conjugative elements (ICEs) play a vital role in bacterial evolution by carrying essential genes that confer adaptive functions to the host. Despite their importance, the mechanism underlying the stable inheritance of ICEs, which is necessary for the acquisition of new traits in bacteria, remains poorly understood. Here, we identified SezAT, a type II toxin-antitoxin (TA) system, and AbiE, a type IV TA system encoded within the ICESsuHN105, coordinately promote ICE stabilization and mediate multidrug resistance in Streptococcus suis. Deletion of SezAT or AbiE did not affect the strain's antibiotic susceptibility, but their duple deletion increased susceptibility, mainly mediated by the antitoxins SezA and AbiEi. Further studies have revealed that SezA and AbiEi affect the genetic stability of ICESsuHN105 by moderating the excision and extrachromosomal copy number, consequently affecting the antibiotic resistance conferred by ICE. The DNA-binding proteins AbiEi and SezA, which bind palindromic sequences in the promoter, coordinately modulate ICE excision and extracellular copy number by binding to sequences in the origin-of-transfer (oriT) and the attL sites, respectively. Furthermore, AbiEi negatively regulates the transcription of SezAT by binding directly to its promoter, optimizing the coordinate network of SezAT and AbiE in maintaining ICESsuHN105 stability. Importantly, SezAT and AbiE are widespread and conserved in ICEs harbouring diverse drug-resistance genes, and their coordinated effects in promoting ICE stability and mediating drug resistance may be broadly applicable to other ICEs. Altogether, our study uncovers the TA system's role in maintaining the genetic stability of ICE and offers potential targets for overcoming the dissemination and evolution of drug resistance. Author summary: Integrative and conjugative elements (ICEs) are mobile genetic elements that play a crucial role in driving bacterial evolution and the acquisition of new traits, including multidrug resistance. An essential prerequisite for bacteria to acquire new features is the stable inheritance of ICEs. In this study, we demonstrated that the type II TA system SezAT and the type IV TA system AbiE play a coordinated role in the genetic stabilization of ICESsuHN105. SezAT and AbiE coordinately promote multidrug resistance in S. suis, an effect mediated by the antitoxins. Our findings suggest that this cooperative mechanism may have broader implications for other ICEs. These revelations contribute to our comprehension of the physiological role of the TA systems and the genetic stabilization mechanisms of ICEs. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results