Intelligent Agents (in English):Agents, Mechanisms, and Collaboration / Agent Perception: Language and Vision / Actions Planning, Reinforcement-Learning / Causality (40 Vorlesungen, Universität zu Lübeck 2019-2024)
o Excursion Fourier Analysis (Proof of Arrows Theorem)
Perception of Agents (Language and Vision)
o Information retrieval and web-mining agents
o Probabilistic dimension reduction, latent content descriptions, topic models, LDA
o Representation learning for sequential structures, embedding spaces, word2vec, CBOW, skip-gram, hierarchical softmax, negative sampling
o Multi-relational latent semantic analysis
o Language models (1d-CNNs, RNNs, LSTMs, ELMo, Transformers, BERT, GPT-3, T5, and beyond), natural language inference and query answering, reinforcement learning (InstructGPT), in-context learning, ChatGPT
o LaMDA, Grounding, embedding knowledge graphs into language models, GNNs
o Vision and Language (2D-CNNs: AlexNet, ResNet / Transfer Learning / ViLBERT / CLIP)
o Generative vision models (DALL-E and beyond) VQ-VAE/d-VAE, DALL-E's transformer
o Summary: Agents and Perception
Actions of Agents (Planning, Causality, and Reinforcement Learning)
o Introduction (Agent, planning, and acting)
o Deterministic (State-variable representation, forward state-space search, heuristic functions, backward search, and plan-space search)
o Temporal (Temporal representation, planning with temporal models, constraint management, and acting with temporal models)
o Nondeterministic (Planning problem, and/or graph search, determination, and online approaches)
o Probabilistic (Stochastic shortest-path problems, heuristic search algorithms, and online approaches including reinforcement learning)
o Decision Making - Foundations (Utility theory, Markov decision processes, and reinforcement learning) - Extensions (Partially observable MDPs and decentralised POMDPs) - Structure (Lifted DecPOMDPs, Factored MDPs, and first-order MDPs)
o Human-aware (Mental models, interpretable behaviour, and explanations)
Causality
o Introduction to causal representations, causal leanring