Report title一:Data confirmation and traceability for large models
Reporting time and place:2024Year9month12Day, PM15:00, School Management Building304Room
Report Summary:
In recent years,With the rapid development of artificial intelligence technology,The application scope of large models in various fields is rapidly expanding。From natural language processing、Image recognition to autonomous driving and medical diagnosis,Large models have shown great potential in improving system intelligence and solving complex problems。All thanks to scale、High quality training data,And advanced computing power support。However,Training data often contains a large amount of personal and company copyright data,Model training also consumes the expensive computing power of large model companies。On the other hand,The powerful output capabilities of large models have also caused many social problems。Therefore,How to prevent data copyright、Model Copyright、And the abuse of model capabilities is critical for trusted deployment of large models。This report will focus on confirming the authority of model training data、Model parameter weight confirmation、And model generated content traceability,Share the team’s recent Bet365 games work on large model authority confirmation and traceability。Through these discussions,Hope to provide valuable insights to researchers and businesses,Helping them better oversee data and models in the development and application of large models,Ensuring safe and ethical use of technology。
About the speaker:
Zhang Zhikun,Currently a researcher in the Hundred Talents Program of the School of Computer Science and Technology, Zhejiang University,Doctoral Supervisor,2023Selected into the National Youth Talent Program (Overseas)。Before joining Zhejiang University,He served as a visiting assistant professor at Stanford University in the United States,GermanyCISPAHead of the Research Group of the Helmholtz Center for Information Bet365 games,Doctoral Supervisor。Before this,He received a PhD from Zhejiang University,And inCISPAEngaged in postdoctoral research,He was also a visiting scholar at Purdue University in the United States。His research interests include Trustworthy Artificial Intelligence、Data Bet365 games and Privacy、and differential privacy。Zhikun published in top conferences and journals40More related papers, including26ArticleCCF-APapers and19Papers from the four top information Bet365 games conferences. He has wonNDSS 2024Outstanding Paper Award (corresponding author),Zhejiang University and Zhejiang Province Outstanding Doctoral Thesis Nomination Award,Second place in the world in the Differential Privacy Synthetic Dataset Challenge held by the National Institute of Standards and Technology。He also serves as severalCCF-Aof the top conferenceTPCMembers includingUSENIX Bet365 games、ACM CCS、NDSS、WWW、KDDandICDE。
Report title二:Bet365 games on public key cryptography in strong adversarial environment
Reporting time and place:2024Year9month12Day, PM15:00, School Management Building304Room
Report Summary:2013Year“Snowden incident”Confirmed that cryptosystems face real-world backdoor attacks。Different from traditional attacks,Password backdoor attack is highly concealed、Destructive,It is mainly through clever tampering with the implementation process of the cryptographic algorithm,Covertly breaching system Bet365 games without attracting user attention。Affected by this incident,The international cryptography community is formed“Cryptography in the post-Snowden era”(post-snowden cryptography) research boom,Mainly studies password Bet365 games in strong adversarial environments,Specific include: (1) Analysis of attack principle; (2) Protection method design。This report will introduce the Bet365 games progress of our Bet365 games group in this field,And put forward prospects for its development trend。
Personal introduction:
Chen Rongmao,Associate Researcher at the School of Computer Science, National University of Defense Technology,Doctoral Supervisor,Host projects such as the National Natural Science Foundation of China Outstanding Youth Fund and the China Association for Science and Technology Youth Care Project,Long-term engagement in public key cryptography theory and application Bet365 games,Have done some original work in password backdoor attack and defense,Guide students to obtain outstanding doctoral dissertations from the Chinese Cryptozoology Society、Excellent doctoral dissertations and international conferences in Hunan ProvinceACASPBest Student Paper Award, etc., served asIEEE TDSC/JCSTEditorial Board andASIACRYPT/ACM CCS/PKCMember of the Program Committee of other important international conferences,Selected for high-level talent project in special fields。
Report title三:IoT-aware Bet365 games and new application practices
Reporting time and place:2024Year9month12Day, PM15:00, School Management Building304Room
Report Summary:
Information perception is the cornerstone of the Internet of Things information system,Smart terminal devices as small as individual users,Dadao Lighthouse Factory、IoT systems in key national areas such as digital transportation,All rely on the sensor mounted on the final node to collect audio、Video、Environment、Interaction and other fine sensory data,To support complex upper-layer application services。With the widespread application of physical network information systems in industrial and daily life scenarios,Their demand for system Bet365 games is also gradually increasing。However,Due to lack of proper Bet365 games protection,Attacks against IoT sensing systems show extremely high diversity。This report will take smart terminal equipment as a case,Safety collection from sensory information、Three aspects of privacy protection of sensory information and new application practice of sensory information,Share the latest research results and research results。
Personal introduction:
Jin Wenqiang,Professor,National Overseas High-Level Young Talents,Hunan Province Hundred Talents Plan Young Talents,Xiaomi Young Scholar,Jiusan Society Central Science Popularization Committee,Deputy Director of the Department of Cyberspace Bet365 games,Director of Information Bet365 games Research Institute of Hunan University,XXXChief Scientist of Innovation Workstation,Expert from Hunan Provincial Network Bet365 games and Informatization Advisory Committee;Main research areas include Internet of Things Bet365 games、Side channel information analysis and utilization、Data privacy protection, etc.;Host the National Natural Science Foundation of China project2Item, Hunan Provincial Key R&D Plan1item、China UnionPay、Multiple scientific Bet365 games projects of Hunan Electric Power Bet365 games Institute and other school-enterprise cooperation,atMobiCom、CCS、NDSS、Usenix Bet365 games、ACSAC、IEEE TMC、TIFS、TDSCPublished multiple high-level papers in top domestic and foreign journals and conferences,Organization establishmentSenSecLabResearch team, awarded "Huawei Terminal Bet365 games" by Huawei2022The title of “Excellent Partner of the Year”.