Search this site
Embedded Files
Towards AGI
  • Home
  • Schedule
  • Projects
  • Topics&Papers
    • Adversarial Robustness
    • Alignment and Safety
    • CompPsych-FoMo
    • Compression and Fast Inference
    • Continual Learning at Scale
    • Emergence & Phase Transitions in ML
    • Foundation Models
    • Generalization (iid and ood)
    • High Performance Computing
    • Knowledge Fusion
    • Neural Scaling Laws
    • Out-of-Distribution Generalization
    • Scaling Laws in Nature
    • State Space Models
    • Time Series Foundation Models
  • Reading Group
Towards AGI
  • Home
  • Schedule
  • Projects
  • Topics&Papers
    • Adversarial Robustness
    • Alignment and Safety
    • CompPsych-FoMo
    • Compression and Fast Inference
    • Continual Learning at Scale
    • Emergence & Phase Transitions in ML
    • Foundation Models
    • Generalization (iid and ood)
    • High Performance Computing
    • Knowledge Fusion
    • Neural Scaling Laws
    • Out-of-Distribution Generalization
    • Scaling Laws in Nature
    • State Space Models
    • Time Series Foundation Models
  • Reading Group
  • More
    • Home
    • Schedule
    • Projects
    • Topics&Papers
      • Adversarial Robustness
      • Alignment and Safety
      • CompPsych-FoMo
      • Compression and Fast Inference
      • Continual Learning at Scale
      • Emergence & Phase Transitions in ML
      • Foundation Models
      • Generalization (iid and ood)
      • High Performance Computing
      • Knowledge Fusion
      • Neural Scaling Laws
      • Out-of-Distribution Generalization
      • Scaling Laws in Nature
      • State Space Models
      • Time Series Foundation Models
    • Reading Group

Schedule

Topics & Papers Reading Group Scaling Workshops    

Paper presentations and projects: schedule & sign up sheet


Deadlines:
Fri, March 22 - project proposals due

Mon, May 6 - final project submission due

Class 1 (Mon, Jan 8, 4:30-6:30pm)

Lecturer: Irina Rish 

Topic: The Large-scale AI Revolution: Breakthroughs and Challenges    (slides)

Papers:  The Bitter Lesson,   GPT-3 paper: Language Models are Few-Shot Learners

Class 2  (Wed, Jan 10, 4:30-6:30pm)

Lecturer: Irina Rish 

Topic:  The Large-scale AI Revolution: Breakthroughs and Challenges  - continued   (slides, video)

Papers:    Training compute-optimal large language  models (summary/blog: New Scaling Laws for LLMs),  Broken Neural Scaling Laws 

CERC-AAI Lab projects on Foundation Models  - please join us!

Class 3 (Mon, Jan 15, 4:30-6:30pm)

Part 1: Lecturer: Irina Rish 

Topic: Introduction to Continual Learning (slides)

Papers:  Continual T0,   Effect of scale on catastrophic forgetting in neural networks  (summary: Effects of Model and Prior Learning Scale on Catastrophic Forgetting),   Foundational Models for Continual Learning: An Empirical Study of Latent Replay


Part 2: Lecturer: Irina Rish 

Topic: Continual Learning at Scale
Papers:  Continual Pre-Training of Large Language Models: How to (re)warm your model.  

Blog: Continual Learning of Foundation Models 

Class 4  (Wed, Jan 17, 4:30-6:30pm)

Part 1: Lecturers: Mohammad Reza Samsami  and Arjun Ashok

Topic:  State-space Models for Time-series: Call for Collaborations  (slides, video)

Papers: Efficiently Modeling Long Sequences with Structured State Spaces
HiPPO: Recurrent Memory with Optimal Polynomial Projections
Combining Recurrent, Convolutional, and Continuous-time Models with Linear State-Space Layers


Part 2: Lecturer: Irina Rish  (video)

Announcements:  Neural Scaling Laws workshops ,  reading group on Scaling Laws & Emergent Behaviors,  Paper presentations and projects: schedule & sign up sheet

Brief discussion on AI Alignment: AI and the paperclip problem,  AI alignment research links, Unsolved Problems in ML Safety,  Concrete Problems in AI Safety,   Alignment Workshop, Introduction to AI Safety, Ethics and Society.

Brief discussion on objective ethincs: Derek Parfit ,  On What Matters (vol 1, vol 2, vol 3). See also Why Anything? Why This?    Reasons and Persons

Class 5  (Mon, Jan 22,  4:30-6:30pm)

Lecturer: Irina Rish 

Topic: Phase Transitions in AI &Emergent Behaviors in Large-Scale models (slides, video)

Tutorial on Phase Transitions (by Guillaume Dumas, at  the 2nd Workshop on Neural Scaling Laws)

Talk by Pascal Jr. Tikeng Notsawo (University of Montreal/Mila): Is grokking predictable?  video Predicting Grokking Long Before it Happens: A look into the loss landscape of models which grok 

The Law of Robustness by Sebastien Bubeck

Papers:  Hard and Easy Distributions of SAT Problems,  Every Monotone Graph Property Has a Sharp Threshold,  Approximability of probability distributions,   Grokking: Generalization Beyond Overfitting on Small Algorithmic Datasets,   GPT-3 paper: Language Models are Few-Shot Learners,   A universal law of robustness via isoperimetry

Class 6  (Wed, Jan 24, 4:30-6:30pm)

Topic:  Albert Gu's talk about his paper on S4 (video)

Papers:    Efficiently Modeling Long Sequences with Structured State Spaces, other SSM papers 

Class 6  (Mon, Jan 29, 4:30-6:30pm)

Part 1: Lecturers:  Benjamin Therien and Avery Ryoo

Topic :  Scaling  Laws for Neural Language Models  (slides, video)


Part 2: Training Compute-Optimal Large Language Models  (Chinchilla Explained: video)

Class 7  (Wed, Jan 31, 4:30-6:30pm)

Part 1: Lecturer: Romain Roy

Topic:  Technical Report: Large Language Models can Strategically Deceive their Users when Put Under Pressure  (slides, video)


Part 2: Lecturer: Irina Rish

Topic: Computational Psychology & Foundation Models  papers  and   ComPsy FoMo Workshop   (video)

Class 8  (Mon, Feb 5, 4:30-6:30pm)

Part 1: Lecturers:  Paria Mehrbod and Vaibhav Singh

Topic:  Scaling Laws for Transfer   (slides, video)  Passcode: +U^C8Zfp


Part 2: Lecturer: Irina Rish

Topic:  Computational Psychoilogy and Foundation Models intersection, and related class projects (video)


Papers:  Huang et al (2023) Who is ChatGPT? Benchmarking LLMs' Psychological Portrayal Using PsychoBench
Gupta et al (2023) Investigating the Applicability of Self-Assessment Tests for Personality Measurement of Large Language Models.  

Explore PsychoBench

AI alignment: A comprehensive survey - TBA

Class 9  (Wed, Feb 7,  4:45-6:30pm)

Part 1: Lecturers:   Thomas Jiralerspong and Dragos Secrieru

Topic:  Are emergent abilities of LLMs a Mirage?  (slides, video)   Video passcode: ^43yC+Ay


Part 2: Lecturers:   Maxime Petrenko and Parviz Haggi

Topic:  HiPPO: Recurrent Memory with Optimal Polynomial Projections  (slides, video)


Part 3: Lecturer: Alexis Roger

Topic:  Robin Suite of Open-Source Foundation Models and Alignment: Status Update and Call for Collaborations   (slides, video)

Class 10  (Mon, Feb 12,  4:45-6:30pm)

Part 1: Lecturers:    Tanner Ducharme and Mahmood Hegazy 

Topic:   Chain of Thought Prompting Elicits Reasoning in Large Language Models (slides, video)   Passcode: wpPHJ1H#


Part 2: Lecturer: Moetez Kdayem 

Topic:  Vision Mamba : Efficient Visual Representation Learning with Bidirectional State Space Model
VMamba: Visual State Space Model   (slides, video-part1, video-part2)

Class 11  (Wed, Feb 14,  4:30-6:30pm)

Part 1: Lecturers:     Arjun Ashok, Andrew Williams 

Topic:  ForecastPFN: Synthetically-Trained Zero-Shot Forecasting    (slides, video)    Passcode: .nO6422A


Part 2: Lecturer: Irina Rish  

Topic:    Computational Psychology & Psychological Computation (slides, video)
    Project Ideas: a Discussion  (slides, video)
  Robin Projects: Call for Collaboration

     ComPsych Projects: Call for Collaboration

Class 12  (Mon, Feb 19,  4:30-6:30pm)

Part 1: Lecturers:   Istabrak Abbes, Megh Thakkar

Topic: Wide Neural Networks Forget Less Catastrophically     (slides, video)    


Part 2: Lecturer: Paloma Fernandez, Anthony Gosselin  

Topic:   Unsolved Problems in ML Safety  (slides, video) 

Class 13  (Wed, Feb  21  4:30-6:30pm)

Part 1: Lecturers:    Emiliano Penaloza  and Sophie Wu

Topic:     Direct Preference Optimization: Your Language Model is Secretly a Reward Model  (slides, video)    


Part 2:  a brief discussion, then youtube talk by Natasha Jaques 

Topic:    RLHF: How to Learn from Human Feedback with Reinforcement Learning   

Class 14  (Mon, Feb  26  4:30-6:30pm)

Part 1: Lecturers:    William Callaghan  and Humza Wajid Hameed 

Topic:    Artificial Intelligence, Values, and Alignment   (slides, video - 1st hr)    


Part 2:  Lecturers: Moksh Jain and Vedant Shah  (slides, video - 2nd hr)    

Topic:   Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback 

Class 13  (Mon, March 4,  4:30-6:30pm)

Part 1: Lecturers:  Jonathan Siu Chi Lim  and  Mina Beiramy 

Topic:  A Generalist Agent (Gato)  (slides, video)


Part 2:  Lecturers:   Xiaoyin Chen and Xnyu Yuan

Topic:   Uncertainty of Thoughts: Uncertainty-Aware Planning Enhances Information Seeking in Large Language Models  (slides, video)

Class 14  (Wed, March 6,  4:30-6:30pm)

Part 1: Lecturers:  Shruti Joshi  and  Aniket Didolkar  

Topic:  Mamba: Linear-Time Sequence Modeling with Selective State Spaces (slides, video - part 1)


Part 2:  Lecturers:   Anirudh  and Dhuruva Priyan 

Topic:   Optimal neural representation in neural systems at the edge of chaos  (slides, video - part 2)

Scaling & Emergence Reading Group  (Mon, March 11, 2pm -3pm)

Lecturer: Irina Rish

Topic: This class is combined with the Scaling & Emergent Phenomena  Reading Group   


Part 1: Lecturers:   Mahsan Abdoli   and Prince Immanuel 

Topic:    Constitutional AI: Harmlessness from AI Feedback (slides, video)    


Class 15  (Mon, March 11,  4:30-6:30pm)

Part 1: Lecturers:  Misha Barth  and  Sammy Sharief 

Topic:  GENERALIZATION IN DIFFUSION MODELS ARISES FROM GEOMETRY-ADAPTIVE HARMONIC REPRESENTATION  ON MEMORIZATION IN DIFFUSION MODELS (slides, video)


Part 2:  Lecturers:   Yuchen Hui and Yicong Li  (slides, video)

Topic:   Mixtral of Experts 

Class 16  (Wed, March 13,  4:30-6:30pm)

Part 1: Lecturers:  Paolo Cudrano

Topic:  Effect Of Model And Pretraining Scale On Catastrophic Forgetting In Neural Networks (slides, video)


Part 2:  Lecturers:   Mahsan Abdoli, Prince Immanuel

Topic:   Constitutional AI: Harmlessness from AI Feedback (slides, video) 

Class 17  (Mon, March 18,  4:30-6:30pm)

Part 1: Lecturers: Nicolas Bernier

Topic: Scaling Data-Constrained Language Models (slides, video)

Class 18  (Wed, March 20,  4:30-6:30pm)

Part 1: Lecturers: Jon Pilarte Bilbao, Karoline Lippert

Topic:  The power of quantum neural networks (slides, video)


Part 2:  Lecturers:    Aloys Portafaix, Congshu Zou

Topic:   UniT: Multimodal Multitask Learning with a Unified Transformer (slides, video)

Class 19  (Mon, March 25,  4:30-6:30pm)

Part 1: Lecturers: Aniket Saxena, Venkatesh Ramesh

Topic:  Retrieval-Augmented Multimodal Language Modeling (slides, video)


Part 2:  Lecturers:   Jeremy Qin, Jinghan Sun

Topic:   In deep reinforcement learning, a pruned network is a good network (slides, video - part 2)

Class 20  (Wed, March 27,  4:30-6:30pm)

Part 1: Lecturers: Darsh Kaushik, Xavier Morin

Topic:  LoRA: Low-Rank Adaptation of Large Language Models (slides, video)


Part 2:  Lecturers:    Jerome Francis, Raj Ghughare

Topic:   Beyond neural scaling laws: beating power law scaling via data pruning (slides, video)

Class 21  (Mon, Apr 1,  4:30-6:30pm)

Part 1: Lecturers: 

Topic: 


Part 2:  Lecturers:  

Topic:   

Class 22  (Wed, Apr 3,  4:30-6:30pm)

Part 1: Lecturers: Meng Cao

Topic:  The Quantization Model of Neural Scaling (slides, video)


Class 23  (Mon, Apr 8,  4:30-6:30pm)

Part 1: Lecturers: Rishika Bhagwatkar, Shubham Gupta

Topic:  Perceiver: General Perception with Iterative Attention (slides, video)


Class 24  (Wed, Apr 10,  4:30-6:30pm)

Part 1: Lecturers: Andrei Mircea

Topic:  Maximal update parameterization / transfer (slides, video)


Part 2:  Lecturers:    Adel Nabli, Johan Obando

Topic:   DiLoCo: Distributed Low-Communication Training of Language Models (slides, video)

Class 25  (Mon, Apr 15,  4:30-6:30pm)

Part 1: Lecturers: Maxime Gevers

Topic:  Segment Anything (slides, video)


Part 2:  Lecturers:    Nithya Shikarpur, Subhrajyoti Dasgupta

Topic:   Adding Conditional Control to Text-to-Image Diffusion Models (slides, video)

Class 26  (Wed, Apr 17,  4:30-6:30pm)

Part 1: Lecturers: Juan David Vargas

Topic:  Recurrent Independent Mechanisms (slides, video)


Class 27  (Mon, Apr 22,  4:30-6:30pm)

Part 1: Lecturers: Sarvjeet Singh Ghotra

Topic:  Video PreTraining (VPT): Learning to Act by Watching Unlabeled Online Videos (slides, video)


Part 2:  Lecturers:    Sparsha Mishra, Maxime Gevers

Topic:   When to Make Exceptions: Exploring Language Models as Accounts of Human Moral Judgment (slides, video)

Class 28  (Wed, Apr 24,  4:30-6:30pm)

Part 1: Lecturers: Neeraj Kumar, Paul Janson


Topic:  Würstchen: An Efficient Architecture for Large-Scale Text-to-Image Diffusion Models (slides, video)


Part 2:  Lecturers:    Ayoub Echchahed, Nassim El Massaudi


Topic:   Stop Regressing: Training Value Functions via Classification for Scalable Deep RL (slides, video)

Course Project Presentations: Mon, Apr 29 - Fri, May 3

Google Sites
Report abuse
Page details
Page updated
Google Sites
Report abuse