On the Relationship Between the Choice of Representation and In-Context Learning
arxiv.org·1d
🧠Machine Learning
YouTube gets ~5% CTR lift on Shorts by replacing embedding tables with Semantic IDs
shaped.ai·1d
📊Feed Optimization
A gentle introduction to Generative AI: Historical perspective
medium.com·4h·
Discuss: Hacker News
🧠Learned Codecs
GNN Blind Spots: The Hidden Cost of Powerful Graph Models
dev.to·3h·
Discuss: DEV
🕸️Graph Embeddings
Why Low-Precision Transformer Training Fails: An Analysis on Flash Attention
huggingface.co·1d·
Discuss: Hacker News
🧠Intelligence Compression
LogSTOP: Temporal Scores over Prediction Sequences for Matching and Retrieval
arxiv.org·2d
🧠Learned Codecs
NExF: Learning Neural Exposure Fields for View Synthesis
m-niemeyer.github.io·22h·
Discuss: Hacker News
🧠Neural Codecs
Doing Math with Embeddings for Better AI Ad Targeting
ethicalads.io·2d·
Discuss: Hacker News
📊Feed Optimization
To Sink or Not to Sink: Visual Information Pathways in Large Vision-Language Models
arxiv.org·1d
🤖Advanced OCR
Contrastive Weak-to-strong Generalization
arxiv.org·1d
⧗Information Bottleneck
RND1: Simple, Scalable AR-to-Diffusion Conversion
radicalnumerics.ai·1d·
Discuss: Hacker News
💻Local LLMs
MARC: Memory-Augmented RL Token Compression for Efficient Video Understanding
arxiv.org·1d
🧠Learned Codecs
SliceFine: The Universal Winning-Slice Hypothesis for Pretrained Networks
arxiv.org·1d
🧠Neural Codecs
CIR-CoT: Towards Interpretable Composed Image Retrieval via End-to-End Chain-of-Thought Reasoning
arxiv.org·1d
🧮Vector Embeddings
Optimal Stopping in Latent Diffusion Models
arxiv.org·1d
🧠Machine Learning
In-Depth Analysis: "Attention Is All You Need"
dev.to·13h·
Discuss: DEV
🧠Intelligence Compression
Show HN: 1M retail interior image dataset for computer vision (UK/US/EU)
groceryinsight.com·17h·
Discuss: Hacker News
🏺Compression Museums
TransFIRA: Transfer Learning for Face Image Recognizability Assessment
arxiv.org·2d
🏛Digital humanities
Gaze on the Prize: Shaping Visual Attention with Return-Guided Contrastive Learning
arxiv.org·1d
🧭Content Discovery
Improving Temporal Understanding Logic Consistency in Video-Language Models via Attention Enhancement
arxiv.org·1d
⏱️Interval Parsing