Research

Publications by capability

100+ peer-reviewed papers spanning goal-driven autonomy, metacognition, and adaptive planning. Dannenhauer & Molineaux, 2010–2025.

100+
Publications
$30M+
Prior Research Funding
25+ yrs
Building Autonomous Decision Making

World Models 10 papers

Delegance businesses use specialized, human-verified, industry-specific knowledge in our world models that governs how a business reacts to new customers and emerging problems. Decisions based on our world models are hallucination-proof and allow Delegance businesses to respond to novel problems in novel ways, rather than following the crowd by using the same LLM advice as their competitors. For operations, Delegance businesses use foundation model technology to create content at AI speeds, with the guidance of robust world model predictions.

Hypothesis Generation 15 papers

A key Delegance capability is to consider alternative hypotheses that underlie observations. This could mean, for example, hypothesizing whether a company is losing customers due to competitors, market downturns, or bad word-of-mouth. We consider multiple possible unknown events grounded in a cultivated world model, and consider evidence for and against each, to assign probability to each hypothesis. This gives Delegance companies the ability to respond to uncertainty without overcommitting.

Human in the loop novelty generationD. Dannenhauer, M. Bercasio, A. Wong 2023 Self-directed learning of action models using exploratory planningD. Dannenhauer, M. Molineaux, M.W. Floyd, N. Reifsnyder, D.W. Aha ACS 2022 Is everything going according to plan? Expectations in goal reasoning agentsD. Dannenhauer, H. Munoz-Avila, N. Reifsnyder AAAI 2019 Learning from exploration: Towards an explainable goal reasoning agentD. Dannenhauer, M.W. Floyd, M. Molineaux, D.W. Aha 2018 Explainable goal reasoningD. Dannenhauer, M.W. Floyd, M. Molineaux, D.W. Aha 2018 An illustrated situation calculus abstraction for iterative explanatory diagnosisC. Task, M.A. Wilson, M. Molineaux, D.W. Aha AI Communications 2018 Towards explainable NPCs: A relational exploration learning agentM. Molineaux, D. Dannenhauer, D.W. Aha 2018 Explaining rebel behavior in goal reasoning agentsD. Dannenhauer, M.W. Floyd, D. Magazzeni, D.W. Aha 2018 Toward problem recognition, explanation and goal formulationS. Kondrakunta, V.R. Gogineni, M. Molineaux, H. Munoz-Avila, M. Oxenham, M.T. Cox Goal Reasoning Workshop 2018 Understanding what may have happened in dynamic, partially observable environmentsM. Molineaux Dissertation 2017 Informed expectations to guide GDA agents in partially observable environmentsD. Dannenhauer, H. Munoz-Avila, M.T. Cox IJCAI 2016 Raising expectations in GDA agents acting in dynamic environmentsD. Dannenhauer, H. Munoz-Avila IJCAI 2015 Continuous explanation generation in a multi-agent domainM. Molineaux, D.W. Aha ACS 2015 Learning unknown event modelsM. Molineaux, D.W. Aha AAAI 2014 DiscoverHistory: Understanding the past in planning and executionM. Molineaux, U. Kuter, M. Klenk AAMAS 2012

Self-Correction 13 papers

Mistakes happen to everyone, even Delegance businesses. What distinguishes us is the way we respond to failures. Delegance businesses use reflective technology to recognize and learn from failures rather than repeating them. This allows us to respond at AI speeds in the face of changing demands, while not succumbing to the ungoverned drift and snowballing errors of systems that depend solely on foundation models for their intelligence.

Our Team

Meet the researchers behind Delegance.

View team
Join the team

Help us build the future of business.

View open roles