[[
wikihub
]]
Search
⌘K
Explore
People
For Agents
Sign in
Explore
People
For Agents
Sign in
@jeremynixon / Thinking / machine-intelligence-research-frontier.md
Suggest edit
Cancel
Submit suggestion
Title
Name
Note
--- title: "Machine Intelligence Research Frontier" visibility: public --- # Machine Intelligence Research Frontier Category: [[machine-intelligence|Machine Intelligence]] [Read the original document](https://docs.google.com/document/d/1G-ppYPhrAm82PqJWidx75EjvMDz8SRLvbC9QXuIq17k/edit?usp=drivesdk&sa=D&ust=1596495076461000&usg=AOvVaw2K-di8EbvAvT4bFLRQ2mPu) <!-- gdoc-inlined --> --- Let’s start with the intersection of the Brain, Deepmind, and OpenAI research frontiers, and follow up with the workshops at ICML, NeurIPS, ICLR, EMNLP and ACL for the last 2 years. Later, I can add CVPR, AAAI. * ICML 2019, ICML 2018 * NeurIPS 2019, NIPS 2018 * ICLR 2019, ICLR 2018 * EMNLP 2018 * ACL 2018 * ICML 2020 * Deepmind Research Overview * Google Brain Research Overview * Open AI Research Overview 1. Transfer Learning / Domain Adaptation 2. Tools, Environment & Datasets 3. Reinforcement Learning 1. Model-based RL[a][b] 2. Exploration in RL[c][d] 3. Multi-Task Learning 4. Imitation Learning +1 4. Safety[e] 5. Deep Learning 1. Convolutional Neural Networks 2. Sequence 39.Modeling 1. Recurrent Neural Networks 2. Attention 3. Scalability and Speed 4. Parallel and Distributed Learning 5. Distillation / Compactness 6. Natural Language Processing & Understanding[f] 1. Word, Phrase, Paragraph, Document Representation 2. Semantics 3. Multilingual Methods 4. Information Extraction 7. Regularization 8. Multi-Modal[g] 9. Generative Models 1. GANs 2. VAEs 3. Normalizing Flows 10. Variational Inference 11. Linear Models 12. Unsupervised Learning 1. Clustering 2. Dimensionality Reduction 3. Autoencoders 13. Representation Learning 14. Memory 15. Multi-Agent Systems[h] 16. Metalearning 1. Neural Programming[i] 2. Hyperparameter Optimization 3. Loss Function Learning[j] 17. Evolution 18. Game Theory 19. General Machine Learning 20. Theory 1. Mean Field Theory 2. Infinite Width NNs 3. Probably Approximate Correctness 4. VC Dimension 21. Neuroscience 22. Interpretability[k] 23. Adversarial Examples 24. Kernel Machines 25. Collaborative Filtering 26. Graphical / Relational Learning 27. Optimization 1. Convex 2. Non-Convex 28. Bandits 1. Multi-Armed Bandit 29. Topological Data Analysis 30. Semi-Supervised Learning 31. Self-Supervised Learning 32. Learning to Rank & Structured Prediction 33. Feature Selection 34. Statistical Machine Learning 1. Bayesian Machine Learning[l] 1. Bayesian Deep Learning 2. Bayesian Nonparametrics 2. Statistical Processes 1. Gaussian Processes 2. Poisson Processes 3. MCMC 4. Uncertainty Estimation 5. Distributional Shift Robustness[m][n] 35. Reasoning 36. Causal Inference[o] 37. Online Learning 38. Active Learning[p] 39. Continual Learning / Life-Long Learning 40. Time Series 41. Information Theory 42. Intuitive Physics 43. Privacy, Anonymity 44. Security[q] 45. Miscellaneous 46. Applications 1. Speech Recognition 2. Image Categorization 3. Image Captioning 4. Natural Language Understanding 1. Machine Translation[r] 2. Language Modeling / Generation 3. Question Answering 4. Summarization 5. Search 6. Parsing 5. Pedestrian Detection 6. Grasp Detection 7. Go 8. Video 9. Dialogue 10. 3D Object Reconstruction 11. Speaker Verification 12. Health Care 13. Theorem Proving 14. Music 15. Pose Estimation 16. Social Media 17. Speech Generation 18. System Design / Device Placement 19. Fairness 20. Super Resolution 21. Chemistry 1. Molecules and Materials 22. Robotics 1. Autonomous Vehicles 23. Physics 24. Games 25. Art 48. New types of hardware: neuromorphic computing or other types could SIGNIFICANTLY speed up some types of AI, with unpredictable consequences Application: 1. Question for each subfield: 2. “If I want to do _____, what concepts do I need to understand, what facts do I need to memorize, what procedures do I need to practice?” 3. Ranking subfields by a standard 1. Impact on safety 2. Impact on general intelligence [a]+ Models of the environment and its dynamics leads to interpretability [b]+1 [c]AIXI theoretical concerns on exploration: "to explore or not to explore" is a safety-related decision, and it's a fundamental uncertainty [d]+1 [e]Of course [f]+ Important for interfacing with the complexity of human goals, values [g]Making ConvNets explain their decisions using natural language and RL agents speak about decision would improve their interpretability (->safety) and how all of them work [h]Important for safety because many "human values" are derived from the need for cooperation in our evolutionary environment. Seems plausible that an AI trained in a multi-agent way would be easier to "control" or at least do positive-sum bargains with. [i]- Program Induction / Source Code Rewrites are likely to lead to unpredictable systems [j]+ Loss function learning / learned objective (for alignment with Human values) [k]+1 [l]Will help with uncertainty [m]Helpful for both inner optimizers, and using a NN to estimate a reward function [n]+1, Uncertainty Estimation is very important for behaving in the real world safely [o]+ Causal graphs of the environment, at the right level of abstraction, gives strong interpretability and controllability of the system [p]Critical for safety - plausibly amount of supervision is the bottleneck for alignment, and active learning seems the most promising way to use supervision more effectively. [q]Very important in short / medium term . Senior people I've talked to say this defending against adversaries (broader definition of security) is actually the #1 problem in ensuring recommender systems are good for users. [r]+ Translation of decisions / models of the world from neuralese into human language --- *Source: [Original Google Doc](https://docs.google.com/document/d/1G-ppYPhrAm82PqJWidx75EjvMDz8SRLvbC9QXuIq17k/edit?usp=drivesdk&sa=D&ust=1596495076461000&usg=AOvVaw2K-di8EbvAvT4bFLRQ2mPu)*