Schedule

Date

Presenters

Topic

Papers

Length

PDF

08/18

Gedas

Course Overview & Transformers

[1] Attention is All You Need

75 min

slides

08/20

Gedas

Vision Transformers

[2] An image is worth 16x16 words: Transformers for image recognition at scale

20 min

slides

Paper Discussion #1

55 min

slides

08/25

Gedas

Data-efficient ViT Training

[3] Training data-efficient image transformers & distillation through attention

20 min

slides

Gedas

[4] Tokens-to-Token ViT: Training Vision Transformers from Scratch on ImageNet

20 min

slides

Paper Battle #1

35 min

slides

08/27

Gedas

Self-attention & Convolutions

[5] Early Convolutions Help Transformers See Better

20 min

slides

Paper Discussion #2

55 min

slides

09/01

TBD

Self-attention & Convolutions

[6] Do Vision Transformers See Like Convolutional Neural Networks?

20 min

slides

Paper Discussion #3

55 min

slides

09/03

TBD

Efficient ViTs

[7] Swin Transformer: Hierarchical Vision Transformer using Shifted Windows

20 min

slides

TBD

[8] SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers

20 min

slides

Paper Battle #2

35 min

slides

09/08

TBD

Scaling ViTs

[9] Scaling Vision Transformers

20 min

slides

Paper Discussion #4

55 min

slides

09/10

TBD

Transformers for Object Detection

[10] End-to-End Object Detection with Transformers

20 min

slides

TBD

[11] Pix2seq: A Language Modeling Framework for Object Detection

20 min

slides

Paper Battle #3

35 min

slides

09/15

TBD

Self-supervised Transformers

[12] Masked Autoencoders Are Scalable Vision Learners

20 min

slides

TBD

[13] Emerging Properties in Self-Supervised Vision Transformers

20 min

slides

Paper Battle #4

35 min

slides

09/17

TBD

Self-supervised Transformers

[14] Multimodal Autoregressive Pre-training of Large Vision Encoders

20 min

slides

Paper Discussion #5

55 min

slides

09/22

Project Proposal (Part 1)

Slides are due 09/21 at 11:59 pm on Canvas.

09/24

Project Proposal (Part 2)

Slides are due 09/23 at 11:59 pm on Canvas.

09/29

TBD

Video Transformers

[15] Is Space-Time Attention All You Need for Video Understanding?

20 min

slides

Paper Discussion #6

55 min

slides

10/01

TBD

Multimodal Transformers

[16] Perceiver: General Perception with Iterative Attention

20 min

slides

TBD

[17] Attention Bottlenecks for Multimodal Fusion

20 min

slides

Paper Battle #5

35 min

slides

10/06

No Class (Well-being Day)

10/08

TBD

Vision-Language Transformers

[18] Improved Baselines with Visual Instruction Tuning

20 min

slides

TBD

[19] CogVLM: Visual Expert for Pretrained Language Models

20 min

slides

Paper Battle #6

35 min

slides

10/13

TBD

Unified Transformers

[20] Janus: Decoupling Visual Encoding for Unified Multimodal Understanding and Generation

20 min

slides

TBD

[21] Chameleon: Mixed-Modal Early-Fusion Foundation Models

20 min

slides

Paper Battle #7

35 min

slides

10/15

No Class (Fall Break)

10/20

Project Milestone (Part 1)

Slides are due 10/19 at 11:59 pm on Canvas.

10/22

Project Milestone (Part 2)

Slides are due 10/21 at 11:59 pm on Canvas.

10/27

TBD

Visual Reasoning

[22] Vision-R1: Incentivizing Reasoning Capability in Multimodal Large Language Models

20 min

slides

TBD

[23] Visual-RFT: Visual Reinforcement Fine-Tuning

20 min

slides

Paper Battle #8

35 min

slides

10/29

TBD

Visual Reasoning

[24] Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?

20 min

slides

Paper Discussion #7

55 min

slides

11/03

TBD

Transformers for 3D Vision

[25] VGGT: Visual Geometry Grounded Transformer

20 min

slides

TBD

[26] Depth Anything 3: Recovering the Visual Space from Any Views

20 min

slides

Paper Battle #9

35 min

slides

11/05

TBD

Transformers for Robotics

[27] OpenVLA: An Open-Source Vision-Language-Action Model

20 min

slides

Paper Discussion #8

55 min

slides

11/10

TBD

Image Generation

[28] MaskGIT: Masked Generative Image Transformer

20 min

slides

TBD

[29] Scalable Diffusion Models with Transformers

20 min

slides

Paper Battle #10

35 min

slides

11/12

TBD

Video Generation

[30] Genie: Generative Interactive Environments

20 min

slides

Paper Discussion #9

55 min

slides

11/17

TBD

Is Attention All You Need?

[31] MetaFormer Is Actually What You Need for Vision

20 min

slides

Paper Discussion #10

55 min

slides

11/19

Final Presentation (Part 1)

Slides are due 11/18 at 11:59 pm on Canvas.

11/24

Final Presentation (Part 2)

Slides are due 11/23 at 11:59 pm on Canvas.

11/27

No Class (Thanksgiving Break)

12/01

No Class

Page updated

Report abuse