https://arxiv.org/abs/2406.06608
Sunday, June 30, 2024
Mermaid Chart, Flowcharts - Basic Syntax
Mermaid Chart, Flowcharts - Basic Syntax
https://mermaid.js.org/syntax/flowchart.html
Summary of AI and Physics Connection
**Summary of AI and Physics Connection:**
Yann LeCun and Derrick Hodge highlight the deep interconnections between AI and physics. LeCun notes that the derivation of backpropagation from classical mechanics and optimal control was detailed in his 1987 PhD thesis and 1988 paper. Hodge elaborates on how various AI methods correspond to physical principles:
1. **Variational Bayesian Inference** relates to thermodynamics.
2. **Bayesian Ensembling** is analogous to path integrals.
3. **Backpropagation** can be derived via the Lagrangian saddle point method.
4. **Convergence in non-convex objectives** mirrors spontaneous symmetry breaking.
5. **ConvNet pooling** parallels renormalization group theory and MERA tensor networks.
6. **Lorentz-equivariant ConvNets**.
These connections suggest that AI and physics share foundational computational mechanisms, leading to unified theories, new research frontiers, and deeper understanding of both fields.
Wednesday, June 26, 2024
TinyML and Efficient Deep Learning Computing
https://hanlab.mit.edu/courses/2023-fall-65940
Efficient AI Computing,
Transforming the Future.
TinyML and Efficient Deep Learning Computing
This course focuses on efficient machine learning and systems. This is a crucial area as deep neural networks demand extraordinary levels of computation, hindering its deployment on everyday devices and burdening the cloud infrastructure. This course introduces efficient AI computing techniques that enable powerful deep learning applications on resource-constrained devices. Topics include model compression, pruning, quantization, neural architecture search, distributed training, data/model parallelism, gradient compression, and on-device fine-tuning. It also introduces application-specific acceleration techniques for large language models and diffusion models. Students will get hands-on experience implementing model compression techniques and deploying large language models (Llama2-7B) on a laptop.
- Live Streaming:https://live.efficientml.ai/
- Time:
Tuesday/Thursday 3:35-5:00pm Eastern Time
- Location:36-156
- Office Hour:
Thursday 5:00-6:00 pm Eastern Time, 38-344 Meeting Room
- Discussion:Piazza
- Homework Submission:Canvas
- Contact:
Instructor
Associate Professor
Announcements
- 2023-12-15
Final project: reports, slides and demo videos
- 2023-12-14
Final report and course evaluation due
- 2023-11-09
Mid-term survey: https://forms.gle/xMgCohDLX73cd4af9
- 2023-10-31
Lab 5 is out.
- 2023-10-19
Schedule
Date
Lecture
Logistics
Sep 7
Lecture
1
:
Introduction
Chapter I: Efficient Inference
Sep 14
Lecture
3
:
Pruning and Sparsity (Part I)
Sep 26
Lecture
6
:
Quantization (Part II)
Sep 28
Lecture
7
:
Neural Architecture Search (Part I)
Lab 1 due (extended to Sep 30 at 11:59 p.m)
Oct 3
Lecture
8
:
Neural Architecture Search (Part II)
Oct 10
Student Holiday — No Class
Oct 17
Lecture
11
:
TinyEngine and Parallel Processing
Chapter II: Domain-Specific Optimization
Oct 24
Lecture
13
:
Transformer and LLM (Part II)
Nov 2
Lecture
16
:
Diffusion Model
Chapter III: Efficient Training
Nov 7
Lecture
17
:
Distributed Training (Part I)
Nov 16
Lecture
20
:
Efficient Fine-tuning and Prompt Engineering
Nov 23
Thanksgiving — No Class
Chapter IV: Advanced Topics
Nov 28
Lecture
22
:
Quantum Machine Learning
Nov 30
Lecture
23
:
Noise Robust Quantum ML
Dec 5
Lecture
24
:
Final Project Presentation
Dec 7
Lecture
25
:
Final Project Presentation
[Slides]
[Video]
[Video (Live)]
Dec 12
Lecture
26
:
Final Project Presentation + Course Summary
[Slides]
[Video]
[Video (Live)]
Dec 14: Project report and course evaluation due
Course Videos
Lecture
1
:
Introduction
Lecture
12
:
Transformer and LLM (Part I)
Lecture
13
:
Transformer and LLM (Part II)
Lecture
16
:
Diffusion Model
Subscribe to:
Posts (Atom)