AI Safety | Kisaco Research

AI Safety

Color: 
#1d3459
Technologist Deep-Dive (Gen AI & Data Science) Track
AI Safety
AI Technologists
Data Science
C-Suite
Moderator

Author:

Sarah Luger

Co-Chair, Data Sets Working Group
MLCommons

Sarah Luger host of the AI Artifacts podcast (www.aiartifacts.net) and the Co-Chair of the Data Sets Working Group for AI benchmarking organization, MLCommons. Data Sets Working Group continues the research initiated with the Rigorous Evaluation of AI Systems workshop series at AAAI Human Computation and AAAI conferences. The goal is to develop robust schemas and infrastructure supporting the Open Source hosting of benchmark evaluation data sets. The group aims to provide free storage for researchers who have human-generated data (spoken word data is the current focus) of generally high quality.

Sarah is a Contributing Member of the MLCommons AI Safety Stakeholder Engagement, Benchmarks and Tests, and Platform Technology working groups. This nonprofit engineering consortium guides the ML industry by developing benchmarks, public datasets, and best practice.

Her current AI Safety work focuses on building LLM Safety Test Sets, Creating Scoring System, and Running Benchmarks. Sarah is leading the subsequent work automating the translation of safety test prompts into in low-resource languages.

Sarah Luger

Co-Chair, Data Sets Working Group
MLCommons

Sarah Luger host of the AI Artifacts podcast (www.aiartifacts.net) and the Co-Chair of the Data Sets Working Group for AI benchmarking organization, MLCommons. Data Sets Working Group continues the research initiated with the Rigorous Evaluation of AI Systems workshop series at AAAI Human Computation and AAAI conferences. The goal is to develop robust schemas and infrastructure supporting the Open Source hosting of benchmark evaluation data sets. The group aims to provide free storage for researchers who have human-generated data (spoken word data is the current focus) of generally high quality.

Sarah is a Contributing Member of the MLCommons AI Safety Stakeholder Engagement, Benchmarks and Tests, and Platform Technology working groups. This nonprofit engineering consortium guides the ML industry by developing benchmarks, public datasets, and best practice.

Her current AI Safety work focuses on building LLM Safety Test Sets, Creating Scoring System, and Running Benchmarks. Sarah is leading the subsequent work automating the translation of safety test prompts into in low-resource languages.

Panelists

Author:

Jonathan Bennion

AI Engineer
Rackspace

Jonathan Bennion

AI Engineer
Rackspace

Author:

Sergey Davidovich

Co-Founder & Chairman
SparkBeyond

Sergey is an entrepreneur, technological visionary and machine intelligence enthusiast, who continually strives to bridge the gap between human and machine reasoning and interaction. He’s passionate about computational knowledge representation, acquisition, storage, reasoning, and processing.
 

Sergey has served in a range of executive technological positions in disruptive startup companies. Prior to co-founding SparkBeyond, Sergey served as GM and SVP of R&D for NewBrandAnalytics, a social business intelligence pioneer. He’s also served as VP R&D of SemantiNet, a semantic reasoning engine, and co-founded Delver, a social search engine that was acquired by Sears, where he served as CTO. Prior to founding Delver, Sergey was the architect of a large-scale award-winning predictive maintenance system.

Sergey Davidovich

Co-Founder & Chairman
SparkBeyond

Sergey is an entrepreneur, technological visionary and machine intelligence enthusiast, who continually strives to bridge the gap between human and machine reasoning and interaction. He’s passionate about computational knowledge representation, acquisition, storage, reasoning, and processing.
 

Sergey has served in a range of executive technological positions in disruptive startup companies. Prior to co-founding SparkBeyond, Sergey served as GM and SVP of R&D for NewBrandAnalytics, a social business intelligence pioneer. He’s also served as VP R&D of SemantiNet, a semantic reasoning engine, and co-founded Delver, a social search engine that was acquired by Sears, where he served as CTO. Prior to founding Delver, Sergey was the architect of a large-scale award-winning predictive maintenance system.

Author:

Vipul Raheja

Applied Research Scientist
Grammarly

Vipul Raheja is an Applied Research Scientist at Grammarly. He works on developing robust and scalable approaches centered around improving the quality of written communication, leveraging Natural Language Processing and Deep Learning. His research interests lie at the intersection of large language models and controllable text generation. He has published several papers at top-tier Machine Learning and Natural Language Processing conferences and is also an organizer of the workshops on Intelligent and Interactive Writing Assistants held at ACL and CHI conferences. He obtained an MS in Computer Science from Columbia University.

Vipul Raheja

Applied Research Scientist
Grammarly

Vipul Raheja is an Applied Research Scientist at Grammarly. He works on developing robust and scalable approaches centered around improving the quality of written communication, leveraging Natural Language Processing and Deep Learning. His research interests lie at the intersection of large language models and controllable text generation. He has published several papers at top-tier Machine Learning and Natural Language Processing conferences and is also an organizer of the workshops on Intelligent and Interactive Writing Assistants held at ACL and CHI conferences. He obtained an MS in Computer Science from Columbia University.