BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:Asia/Seoul BEGIN:STANDARD DTSTART:19881009T020000 TZOFFSETFROM:+1000 TZOFFSETTO:+0900 TZNAME:KST END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20261030T020827Z UID:9A185987-BECD-4E67-944C-6B69F6831BC1 DTSTART;TZID=Asia/Seoul:20261113T150000 DTEND;TZID=Asia/Seoul:20261113T175000 DESCRIPTION:Speaker: Abdul Majeed\, PhD.\, University of Warwick\, England\ n\nSynthetic data is a cornerstone for AI development as it resolves data scarcity and privacy challenges. Synthetic data is artificially generated information that reproduces important patterns and relationships of real d ata without directly copying its records. Fundamental topics include gener ation methods—such as statistical sampling\, simulation/agent-based meth ods\, data augmentation\, and rule-based approaches—along with data type s\, preprocessing\, train–test separation\, and evaluation of realism\, utility\, diversity\, and privacy. Advanced methods include generative adv ersarial networks (GANs)\, variational autoencoders (VAEs)\, diffusion mod els\, large language models\, and domain-specific simulators for producing tabular\, image\, text\, time-series\,\nand graph data. Important advance d concepts include conditional generation\, rare-event synthesis\, fairnes s\, distribution shift\, privacy attacks\, differential privacy\, federate d generation\, and model-based validation. This lecture introduces synthet ic data approaches (rule-based\, statistical-based\, hand-engineered\, etc .)\, generator types (GANs\, GAN variants\, VAEs)\, adversarial learning f rameworks\, and their powerful ability to generate realistic data. Applica tions in vision and synthesis are explored across disciplines (e.g.\, medi cal\, transport\, ambient assisted living\, LLMs\, etc.). Advanced data ge nerators such as vision models\, federated models\, and LLMs will also be discussed in relation to conventional approaches. Lastly\, the challenges related to bias\, fairness\, privacy\, and attack vectors will also be dis cussed in detail. Some discussion around metrics will also be provided to highlight synthetic use across domains and scenarios. Some discussion arou nd both attacks (e.g.\, launching a membership attack) and defense vectors (defense against attribute/membership leakage) of the synthetic data will also be provided. Lastly\, some uses of SD in agentic AI and LLMs will al so be discussed to provide information on the latest trends in this line o f work. Finally\, challenges will be presented through detailed synthesis of state-of-the-art literature to demonstrate where this technology is hea ding\, which will allow participating researchers to pursue the latest top ics/challenges.\n\nAgenda: \nInstructor: Abdul Majeed\, PhD.\, University of Warwick\, England\n\nVirtual: https://events.vtools.ieee.org/m/589864 LOCATION:Virtual: https://events.vtools.ieee.org/m/589864 ORGANIZER:sohwang@gachon.ac.kr SEQUENCE:5 SUMMARY:Synthetic Data Generators: The Deep Dive into Synthetic Realities URL;VALUE=URI:https://events.vtools.ieee.org/m/589864 X-ALT-DESC:Description: <br /><p>Speaker: Abdul Majeed\, PhD.\, University of Warwick\, England</p>\n<p>Synthetic data is a cornerstone for AI develo pment as it resolves data scarcity and privacy challenges. Synthetic data is artificially generated information that reproduces important patterns a nd relationships of real data without directly copying its records. Fundam ental topics include generation methods&mdash\;such as statistical samplin g\, simulation/agent-based methods\, data augmentation\, and rule-based ap proaches&mdash\;along with data types\, preprocessing\, train&ndash\;test separation\, and evaluation of realism\, utility\, diversity\, and privacy . Advanced methods include generative adversarial networks (GANs)\, variat ional autoencoders (VAEs)\, diffusion models\, large language models\, and domain-specific simulators for producing tabular\, image\, text\, time-se ries\,<br>and graph data. Important advanced concepts include conditional generation\, rare-event synthesis\, fairness\, distribution shift\, privac y attacks\, differential privacy\, federated generation\, and model-based validation. This lecture introduces synthetic data approaches (rule-based\ , statistical-based\, hand-engineered\, etc.)\, generator types (GANs\, GA N variants\, VAEs)\, adversarial learning frameworks\, and their powerful ability to generate realistic data. Applications in vision and synthesis a re explored across disciplines (e.g.\, medical\, transport\, ambient assis ted living\, LLMs\, etc.). Advanced data generators such as vision models\ , federated models\, and LLMs will also be discussed in relation to conven tional approaches. Lastly\, the challenges related to bias\, fairness\, pr ivacy\, and attack vectors will also be discussed in detail. Some discussi on around metrics will also be provided to highlight synthetic use across domains and scenarios. Some discussion around both attacks (e.g.\, launchi ng a membership attack) and defense vectors (defense against attribute/mem bership leakage) of the synthetic data will also be provided. Lastly\, som e uses of SD in agentic AI and LLMs will also be discussed to provide info rmation on the latest trends in this line of work. Finally\, challenges wi ll be presented through detailed synthesis of state-of-the-art literature to demonstrate where this technology is heading\, which will allow partici pating researchers to pursue the latest topics/challenges.</p><br /><br /> Agenda: <br /><p>Instructor: Abdul Majeed\, PhD.\, University of Warwick\, England</p> END:VEVENT END:VCALENDAR