{"692854":{"#nid":"692854","#data":{"type":"news","title":"Can a Machine or AI Agent Be Surprised? Helping Autonomous Systems Respond to the Unexpected  ","body":[{"value":"\u003Cp\u003ELet\u2019s say you ask ChatGPT a question that stumps it, or a Waymo vehicle encounters something unusual in the road, or an autonomous factory faces an unexpected disruption.\u003C\/p\u003E\u003Cp\u003EMost people would recognize that something unexpected has happened and adjust accordingly. \u0026nbsp;\u003C\/p\u003E\u003Cp\u003EFor machines, it\u2019s not always that simple. \u0026nbsp;\u003C\/p\u003E\u003Cp\u003EResearchers at \u003Ca href=\u0022https:\/\/www.isye.gatech.edu\/\u0022\u003EGeorgia Tech\u2019s H. Milton Stewart School of Industrial and Systems Engineering (ISyE) \u003C\/a\u003Eare addressing this very challenge: how can autonomous systems recognize when something unexpected has happened, determine whether it matters, and decide how to respond?\u003C\/p\u003E\u003Cp\u003EIn a recent \u003Ca href=\u0022https:\/\/pubsonline.informs.org\/doi\/10.1287\/ijds.2026.0182\u0022\u003Epaper published in the INFORMS Journal on Data Science\u003C\/a\u003E, the research team introduced a new framework called Mutual Information Surprise, designed to help machines identify meaningful surprises in complex environments. The work could help lay the foundation for AI agents, robots, and autonomous systems that can better navigate unexpected situations, avoid costly mistakes, and make more informed decisions when faced with situations they weren\u2019t designed or trained to handle. \u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u0022Traditional automation systems are designed to follow instructions. They do not have a sense of surprise. Machines follow a predetermined recipe and produce the expected result,\u0022 said \u003Ca href=\u0022https:\/\/www.isye.gatech.edu\/users\/yu-ding\u0022\u003EYu Ding, the Anderson-Interface Chair and professor in ISyE\u003C\/a\u003E, who is leading this project. \u0022 Intelligent agents, however, must do more than simply follow instructions. They need to recognize when their current understanding or reasoning process has become inadequate. Humans routinely use surprise for this purpose. An autonomous machine needs an analogous computational capability.\u0022\u003C\/p\u003E\u003Cp\u003EThis research concept was jointly worked out by \u003Ca href=\u0022https:\/\/www.isye.gatech.edu\/users\/xiao-liu\u0022\u003EXiao Liu, the David M. McKenney Family Associate Professor\u003C\/a\u003E, as well as a former postdoctoral fellow in ISyE, Yinsong Wang, and Ph.D. student Quan Zeng.\u003C\/p\u003E\u003Cp\u003EWhile several methods already exist to measure surprise in machines, such as the \u0022Shannon Surprise\u0022 and \u0022Bayesian Surprise,\u0022 the ISyE team saw limitations in both approaches. \u0026nbsp;\u003C\/p\u003E\u003Cp\u003EShannon Surprise tends to focus on the rarity of an event, while Bayesian Surprise focuses on how much a new observation changes a system\u0027s beliefs.\u003C\/p\u003E\u003Cp\u003ETo illustrate the limitations of existing surprise measures, Ding pointed to the examples of someone noticing a specific type of car parked nearby or a student who consistently earns top grades but suddenly performs poorly on an exam.\u003C\/p\u003E\u003Cp\u003EExisting surprise measures don\u0027t always flag the right kinds of moments to be surprised, Ding said. To the examples given, the Shannon Surprise would flag the specific car parked nearby, although people typically wouldn\u2019t be surprised by that event. Conversely, people would be surprised by the student performing poorly, but the Bayesian Surprise would not flag it. \u0026nbsp; \u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u0022We introduce this new surprise definition because it measures the gain of knowledge and epistemic progression, thus capturing the right moments when a system should be surprised and ignoring the moments when it shouldn\u0027t,\u0022 Ding added. \u0026nbsp;\u003C\/p\u003E\u003Cp\u003EThe framework could eventually help systems determine when to continue as planned or when to stop and reassess, whether that\u0027s an AI assistant struggling to answer a question, a robot encountering a situation it has never seen before, or an autonomous factory detecting a disruption that could affect production and require human intervention.\u003C\/p\u003E\u003Cp\u003ELooking ahead, the ISyE research team is exploring potential applications in engineering autonomous systems and is developing a proposal for federally funded programs. \u0026nbsp;\u003C\/p\u003E","summary":"","format":"limited_html"}],"field_subtitle":"","field_summary":[{"value":"\u003Cp\u003E\u003Cem\u003E\u003Cstrong\u003EResearchers at Georgia Tech\u0027s H. Milton Stewart School of Industrial and Systems Engineering (ISyE) are rethinking how intelligent systems recognize and respond to the unexpected.\u0026nbsp;\u003C\/strong\u003E\u003C\/em\u003E\u0026nbsp;\u003C\/p\u003E","format":"limited_html"}],"field_summary_sentence":[{"value":"Researchers at Georgia Tech\u0027s H. Milton Stewart School of Industrial and Systems Engineering (ISyE) are rethinking how intelligent systems recognize and respond to the unexpected.  "}],"uid":"35874","created_gmt":"2026-09-28 14:47:39","changed_gmt":"2026-09-29 12:31:45","author":"Anna Akins","boilerplate_text":"","field_publication":"","field_article_url":"","location":"Atlanta, GA","dateline":{"date":"2026-09-28T00:00:00-04:00","iso_date":"2026-09-28T00:00:00-04:00","tz":"America\/New_York"},"extras":[],"hg_media":{"681269":{"id":"681269","type":"image","title":"The ISyE research team from left to right: Xiao Liu, the David M. McKenney Family Associate Professor; Yu Ding, the Anderson-Interface Chair and professor and project lead; and Ph.D. student Quan Zeng.","body":null,"created":"1790606883","gmt_created":"2026-09-28 14:48:03","changed":"1790606883","gmt_changed":"2026-09-28 14:48:03","alt":"The ISyE research team from left to right: Xiao Liu, the David M. McKenney Family Associate Professor; Yu Ding, the Anderson-Interface Chair and professor and project lead; and Ph.D. student Quan Zeng.","file":{"fid":"265638","name":"IMG_2951--1-.jpg","image_path":"\/sites\/default\/files\/2026\/09\/28\/IMG_2951--1-.jpg","image_full_path":"http:\/\/hg.gatech.edu\/\/sites\/default\/files\/2026\/09\/28\/IMG_2951--1-.jpg","mime":"image\/jpeg","size":875476,"path_740":"http:\/\/hg.gatech.edu\/sites\/default\/files\/styles\/740xx_scale\/public\/2026\/09\/28\/IMG_2951--1-.jpg?itok=sQjT7D4T"}},"681268":{"id":"681268","type":"image","title":"ISyE researchers contributed to the development of a new framework called Mutual Information Surprise to help autonomous systems recognize and respond to meaningful unexpected events in complex environments.","body":null,"created":"1790606883","gmt_created":"2026-09-28 14:48:03","changed":"1790606883","gmt_changed":"2026-09-28 14:48:03","alt":"ISyE researchers developed a new framework called Mutual Information Surprise to help autonomous systems recognize and respond to meaningful unexpected events in complex environments.","file":{"fid":"265637","name":"AI-rendering--1-.png","image_path":"\/sites\/default\/files\/2026\/09\/28\/AI-rendering--1-.png","image_full_path":"http:\/\/hg.gatech.edu\/\/sites\/default\/files\/2026\/09\/28\/AI-rendering--1-.png","mime":"image\/png","size":2531952,"path_740":"http:\/\/hg.gatech.edu\/sites\/default\/files\/styles\/740xx_scale\/public\/2026\/09\/28\/AI-rendering--1-.png?itok=TnYqo2kO"}},"681270":{"id":"681270","type":"image","title":"The research team stands outside of George Tower, ISyE\u0027s new home in Tech Square.","body":null,"created":"1790606883","gmt_created":"2026-09-28 14:48:03","changed":"1790606883","gmt_changed":"2026-09-28 14:48:03","alt":"The research team stands outside of George Tower, ISyE\u0027s new home in Tech Square.","file":{"fid":"265639","name":"IMG_2960--1-.jpg","image_path":"\/sites\/default\/files\/2026\/09\/28\/IMG_2960--1-.jpg","image_full_path":"http:\/\/hg.gatech.edu\/\/sites\/default\/files\/2026\/09\/28\/IMG_2960--1-.jpg","mime":"image\/jpeg","size":3269861,"path_740":"http:\/\/hg.gatech.edu\/sites\/default\/files\/styles\/740xx_scale\/public\/2026\/09\/28\/IMG_2960--1-.jpg?itok=QMax2sXU"}}},"media_ids":["681269","681268","681270"],"groups":[{"id":"1237","name":"College of Engineering"},{"id":"1188","name":"Research Horizons"},{"id":"1242","name":"School of Industrial and Systems Engineering (ISYE)"}],"categories":[{"id":"194606","name":"Artificial Intelligence"},{"id":"145","name":"Engineering"},{"id":"194685","name":"Manufacturing"},{"id":"135","name":"Research"}],"keywords":[],"core_research_areas":[{"id":"193655","name":"Artificial Intelligence at Georgia Tech"},{"id":"39461","name":"Manufacturing, Trade, and Logistics"},{"id":"39541","name":"Systems"}],"news_room_topics":[],"event_categories":[],"invited_audience":[],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[{"value":"\u003Cp\u003EAnna Akins, Communications Manager I\u003Cbr\u003EResearcher photos taken by Medha Gollakoti, ISyE student assistant.\u0026nbsp;\u003Cbr\u003EAI illustration created using Microsoft Copilot.\u0026nbsp;\u003C\/p\u003E","format":"limited_html"}],"email":[],"slides":[],"orientation":[],"userdata":""}}}