69 questions · 15 almanac · 14 findings

Artificial Intelligence

The science and engineering of intelligent machines.

Almanac Building and Doubting the Thinking Machine
  • What did Minsky and Papert's book Perceptrons prove about single-layer neural networks?
  • Why did Perceptrons cool enthusiasm for connectionist AI approaches for a decade?
  • What role did Minsky play in founding the MIT Artificial Intelligence Laboratory?
Almanac Naming a Field That Didn't Exist Yet
  • How did John McCarthy come to coin the term 'artificial intelligence' at the 1955 Dartmouth workshop proposal?
  • What made LISP's design unusual, and why did treating code as data matter for symbolic AI?
  • Why did early AI researchers like McCarthy believe human-level machine intelligence was close at hand?
Almanac Thinking, Defined as Symbol Shuffling
  • What is the physical symbol system hypothesis, and what does it claim about intelligence?
  • How did Newell and Simon's General Problem Solver extend the ideas behind Logic Theorist?
  • In what sense did Newell and Simon argue human problem-solving and computation are the same activity?
Almanac Nobody Actually Optimises Anything
  • Why did Simon argue that the classical model of a fully rational economic actor was unrealistic?
  • How did Simon's work span economics, psychology, and cognitive science at once?
Almanac Teaching Machines to Know a Few Things Well
  • How did Edward Feigenbaum's expert systems like DENDRAL and MYCIN differ from earlier AI research aimed at universal reasoning engines?
  • How accurate was MYCIN at diagnosing blood infections compared to human clinicians?
  • What made real-time continuous speech recognition, which Raj Reddy pioneered at Carnegie Mellon, so difficult in the 1970s and 1980s?
Almanac Chess Was No Longer Ours Alone
  • What upgrades did IBM make to Deep Blue between the 1996 loss and the 1997 rematch?
  • Why did Kasparov suspect human intervention in Deep Blue's second game, and how did IBM respond?
  • What was the final score of the 1997 rematch between Deep Blue and Kasparov?
Finding Advancing Education through Tutoring Systems: A Systematic Literature Review
  • What distinct categories of tutoring systems emerge from a Latent Class Analysis of 86 studies on Intelligent Tutoring Systems and Robot Tutoring Systems?
  • How do Intelligent Tutoring Systems and Robot Tutoring Systems complement each other in delivering personalized, adaptive instruction versus social-emotional engagement?
  • What AI techniques, such as Bayesian Knowledge Tracing and Large Language Models, are driving improvements in adaptability and learning outcomes within tutoring systems?
  • What ethical, scalability, and cognitive-adaptability gaps remain unresolved in current tutoring system research, and how might hybrid ITS-RTS solutions address them?
Finding AI-Enabled Serious Games: Integrating Intelligence and Adaptivity in Training Systems
  • How does this chapter distinguish between instructional intelligence and adaptivity in AI-enabled serious games?
  • What historical trajectory connects early computer-assisted instruction and intelligent tutoring systems to today's AI-enabled adaptive learning architectures?
  • In what ways might large language models, reinforcement learning, and agent-based architectures enable more integrated instructional adaptation within serious games?
  • What practical and research challenges, such as explainability, validation, and computational cost, limit the deployment of AI-enabled serious games despite their promise?
Finding Beyond Answers: Large Language Model-Powered Tutoring System in Physics Education for Deep Learning and Precise Understanding
  • How does Physics-STAR use step-by-step guidance and reflective prompts instead of direct answers to support deeper learning in high school physics?
  • What gains in accuracy and efficiency did students show on conceptual, computational, and informational physics questions after using Physics-STAR?
  • Why did complex information problems see a 100% score improvement while overall efficiency gains were more modest at 5.95%?
Finding Intelligent Tutoring Systems for Generation Z's Addiction
  • What design components define a next-generation intelligent tutoring system that adapts to individual learners?
  • What limitations do conditional-statement-driven dialogue agents have compared to neural network based approaches for tutoring?
  • In what extended reality contexts, such as rehabilitation and trauma healing, have computer-based tutoring agents already been deployed?
Finding Leveraging Deep Reinforcement Learning for Metacognitive Interventions across Intelligent Tutoring Systems
  • How does a deep reinforcement learning agent that adapts to a student's changing metacognitive level compare to a static classifier-based approach for teaching strategy-switching in intelligent tutoring systems?
  • Why do static, classifier-based metacognitive interventions only benefit students who already know how to use the target strategy, while adaptive DRL-based interventions close the skills gap across the whole class?
  • What evidence shows that DRL-based metacognitive interventions in one tutoring domain (a logic tutor) transfer to prepare students for future learning in an unrelated domain (a probability tutor)?
Finding Personalized Multimodal Feedback Using Multiple External Representations: Strategy Profiles and Learning in High School Physics
  • How does elaborated feedback across verbal, graphical, and mathematical representations relate to post-test performance in high school physics?
  • Why does using a diverse set of representations benefit students with lower representational competence more than those with higher competence?
  • What do these findings imply for designing adaptive, multimodal feedback in intelligent tutoring systems that use large language models?
Finding Precision Proactivity: Measuring Cognitive Load in Real-World AI-Assisted Work
  • How can intrinsic and extraneous cognitive load be estimated from transcripts of real-world AI-assisted knowledge work?
  • Why does extraneous cognitive load have a much larger negative association with task performance than intrinsic load in AI-assisted work?
  • How does AI-generated content usage compensate for cognitive load, and why is this compensation only partial?
  • How does professional expertise moderate the relationship between cognitive load and performance benefits from AI assistance?
Finding Review of intelligent tutoring systems using bayesian approach
  • How do model-tracing cognitive tutors represent a student's problem-solving strategies and misconceptions?
  • Why are Bayesian Networks well suited for decision-making in intelligent tutoring systems?
  • In what ways do intelligent tutoring systems adapt instruction to a learner's individual progress?
Finding Spaced Repetition and Retrieval Practice Empowered by AI
  • How have spaced repetition systems evolved from the Leitner box method through the SM-2 algorithm to modern machine-learning approaches like SSP-MMC and LSTM-HLR?
  • In what ways do deep learning models like LSTM-HLR and natural language processing enable AI systems to predict optimal review timing and assess partial knowledge?
Finding The Dual Role of Abstracting over the Irrelevant in Symbolic Explanations: Cognitive Effort vs. Understanding
  • How do removal and clustering, as two distinct abstraction operations over irrelevant details in symbolic explanations, differ in their effects on human understanding versus cognitive effort?
  • How can Answer Set Programming be used to formally define which details in a logical trace are 'irrelevant' and thus eligible for abstraction in an explanation?
  • What does the finding that clustering improves understanding while removal reduces cognitive effort imply for designing human-centered explanations from symbolic AI systems?
Finding The Life Cycle of Large Language Models: A Review of Biases in Education
  • Why do traditional machine learning bias metrics fail to transfer to LLM-generated educational content?
  • At which stages of the LLM life cycle, from development to deployment, can algorithmic bias enter educational applications?
  • How might personalized LLM tutoring affect cognitive load management for different student populations?
  • Why can tailoring an LLM's responses to individual learners be pedagogically desirable rather than a form of unfair bias?