#54: Exploring Intelligent Software with Dick Stottler
Nov 20, 2024
auto_awesome
Dick Stottler, Co-founder and CEO of Stottler Henke, shares his extensive expertise in AI solutions tailored for complex problems. He discusses innovative applications of intelligent software in defense, space missions, and education, emphasizing human oversight in automated systems. Stottler dives into cognitive task analysis in satellite operations and the importance of symbolic techniques in AI to improve accuracy. He also addresses misconceptions about AI's capabilities and explores the evolving public perception of intelligent systems.
Dick Stottler illustrates the significance of cognitive task analysis in designing AI systems that effectively capture human thought processes.
The podcast emphasizes the importance of combining autonomous AI systems with human oversight, particularly in high-stakes environments like space and military applications.
Deep dives
The Evolution of AI in High-Stakes Environments
Dick Stottler discusses his extensive experience in developing artificial intelligence (AI) technologies since 1988, particularly in high-stakes domains such as military and space applications. His company, Stotler-Henke, employs a significant range of AI techniques to create practical solutions that solve complex problems. This includes intelligent tutoring systems for training military personnel and automated planning for NASA missions. Stottler emphasizes the importance of understanding human thought processes, achieving this through cognitive task analysis to better inform the design and function of AI systems.
Case-Based Reasoning and Human Decision-Making
Stottler explains the concept of case-based reasoning, where AI mimics the way humans make decisions based on previous experiences. This method involves recalling past solutions and adapting them to current challenges, similar to how an individual might evaluate the worth of a house by comparing it to others. He illustrates this with examples from manufacturing and firefighting, where AI systems enhance decision-making by applying learned patterns to immediate situations. The focus remains on accurately capturing the nuances of human cognition to improve AI functionalities.
Hybrid Systems and Autonomous Decision-Making
The podcast highlights various projects Stottler's team is currently working on, including a rover for lunar missions that requires autonomous decision-making capabilities. This rover is designed to quickly assess mechanical stability to prevent tipping, an application where human intervention is impossible due to communication delays. Additionally, the team is developing drone systems for wildfire management, employing a mix of human cognitive understanding and advanced AI. Stottler notes that their approach effectively combines both autonomous systems and human oversight to ensure safety in high-risk scenarios.
The Future of AI: Balancing Automation with Human Insight
Stottler shares insights on the shifting perceptions of AI and emphasizes that true AI doesn't mimic human consciousness but serves as a powerful tool under human control. He cautions against misconceptions that generative AI can operate independently, highlighting the need for a balanced approach that incorporates both symbolic processing and machine learning techniques. Stottler advises young AI developers to pursue broad education and curiosity beyond computer science, advocating that knowledge across various domains enhances problem-solving capabilities. Ultimately, successful AI integration requires collaboration between engineers, domain experts, and clear communication of cognitive processes.
Dick co-founded Stottler Henke in 1988 as a software company dedicated to providing practical solutions to difficult problems by skillfully drawing upon a large repertoire of artificial intelligence technologies. Under Dick’s leadership, Stottler Henke has grown steadily and profitably into a 60-person research and software development company with distinctive expertise in automated planning and scheduling, intelligent tutoring systems, intelligent simulation, sensor data fusion, and autonomous systems.
Dick combines a strong applied research record in artificial intelligence with practical experience in rapid and efficient knowledge engineering. He has led the development of intelligent tutoring systems that encode the expertise of instructors to provide practice-based learning and automated evaluation of student performance. Subject areas include navy tactics; army tactics, command, and control; sonar data analysis; astronaut training; helicopter cockpit operations; and battlefield emergency medicine. He led the development of intelligent planning systems for NASA space shuttle missions, satellite communications and sensor scheduling, and aircraft assembly. He also led the development of intelligent systems that autonomously control groups of drones to automatically plan and execute tasks to achieve mission objectives and react to unpredictable events.
Dick has written or presented more than 80 papers and articles for publications such as the proceedings of the International Joint Conference on Artificial Intelligence (IJCAI). He received his BS in engineering from Cornell University and his MS in computer science (artificial intelligence) from Stanford University.