ML Paper Review

My attempt to keep up with the latest development in the AI/ML world. Research area: NLP, Deep learning for Recommender System, Computer vision

Review ItemSAGE Paper

2 minute read

Transformer-based, multi-modality recsys from Pinterest

Review Vision Transformer paper

2 minute read

Proof that Transformer could work with Vision

Review LLaMA paper

2 minute read

Meta released a SOTA LMM for the research community

Review Segment Anything paper

2 minute read

Foundational promptable image segmentation model

Review Generative Adversarial Nets (GAN)

2 minute read

Generative model G + discriminative model D

Review Generative Adversarial Nets (GAN)

1 minute read

Generative model G + discriminative model D

Review Diffusion Models

2 minute read

Advancement for efficient training of Denoising Diffusion models

LORA: LOW-RANK ADAPTATION OF LARGE LANGUAGE MODELS

3 minute read

Parameter-Efficient fine-tuning for LLMs

Contrastive Learning - SimCLR

1 minute read

Basics of contrastive learning

DINO - self-distillation with no labels

1 minute read

Using momentum for the teacher and balancing between centering and sharpening to avoid mode collapse

BERT

1 minute read

Review the Bert paper from Google

Masked Auto Encoder for Vision

2 minute read

Vision’s BERT - Everything is a dog

CLIP

1 minute read

General-purpose Visual-Natural Language aligned encoder

Imagen paper

1 minute read

Photorealistic Text-to-Image Diffusion Models with Deep Language Understanding

Visual Robotic Bridge

1 minute read

Affordances from Human Videos as a Versatile Representation for Robotics

Domain-Specific Batch Norm

2 minute read

Domain-Specific Batch Norm (DSBN) for Unsupervised Adaptation