時間：3月18日上午9:30
地點：新校區新珈樓B103教室
報告題目：淺談論文調研與研究方法
報告人：覃立波教授
報告人簡介：覃立波，中南大學計算機學院特聘教授，主要研究方向為任務型對話系統和自然語言處理。在ACL、EMNLP、AAAI、IJCAI等國際會議上發表論文多篇，研究成果曾入選Paper Digest高影響力論文及獲得EMNLP2022 MMNLU Workshop最佳論文獎。擔任IJCAI2021高級程序會委員（SPC）、EMNLP2022領域主席。
報告摘要：在科研過程中，初入科研的同學往往會有較長的不适應過渡期。對于初學同學來說最重要的是第一個idea的産生。在這個過程中，如何進行調研，如何去有思路的找到創新的問題都是一個極大的挑戰。在這個報告中，我将淺析如何進行一個方向的調研工作，并講解如何逐步剖析去尋找到一個科研問題，希望能幫助初入科研同學快速進入狀态。
報告題目：ChatGPT, using chat to connect human with language models
報告人：王本友助理教授
報告人簡介：Benyou Wang is an assistant professor in the School of Data Science, The Chinese University of Hong Kong, Shenzhen. He was a Marie Curie Researcher at the European Union and got his Ph.D. degree from the University of Padua, Italy in 2022. So far, he and his collaborators have won the Best Paper Nomination Award in SIGIR 2017, Best Explainable NLP Paper in NAACL 2019, and Best Paper in NLPCC 2022. He is committed to building novel, explainable, robust, and efficient natural language processing systems that are with both technical rationality and linguistic motivations.
報告摘要：In just three months, ChatGPT, a groundbreaking language model, has gained worldwide recognition. This talk will introduce the concept of "language models" and "prompts" where prompts serve as an interface connecting human with language models. By using text prompts, one could enjoy the merits of ChatGPT for various features including zero-shot generation, in-context learning, complex reasoning, chain of thought, interactive chat, etc; although some of them have already explored by ChatGPT's earlier versions. We will further discuss new features, technical map, limitations, as well as some case study of ChatGPT. To help you make the most of ChatGPT, we will also provide some basic usage guides that can help improve your productivity. Lastly, we will discuss the possibility of creating a local version of ChatGPT and our work on a medical ChatGPT. We are excited to witness the impact that ChatGPT will have on the industry, as well as the wider scientific and engineering communities.
報告題目：快手推薦系統技術分享--- WWW2023專題報告
報告人：劉殊暢博士
報告人簡介：劉殊暢，快手高級算法工程師，畢業于羅格斯大學，師從張永鋒老師。主要研究方向為推薦系統、遷移學習和端計算，發表CCF-A/B類論文10餘篇。WWW/SIGIR/KDD/IJCAI/AAAI評審委員，TORS期刊評審委員。
報告摘要：1月25日，國際學術會議WWW 2023論文接收結果公布。快手社區科學線有多篇論文被錄用，本次報告将介紹其中快手推薦策略中台組的3篇工作。快手推薦場景相比傳統推薦場景具有大流量、更新快、玩法複雜、用戶頻繁交互的特點，尤其在持續交互和留存優化等長期目标上使用強化學習解決方案時，這些挑戰被進一步放大。報告分享内容将着重探讨其中多目标推薦、強化學習行為空間探索、以及留存優化問題的應對策略與思路。
邀請人：鄒立新副教授、李晨亮教授
