Launch the Telescoop'

The Telescoop' Bubble is dedicated to the AI research papers.

Semantic Scholar API

This Tool has been realized by IDRIS/CNRS Team. It could be implemented due to the excellence of Semantic Scholar API and the visualization Bokeh library.

Please read the paper The Semantic Scholar Open Data Platform to know how Semantic Scholar API is great.



And please take a look at Scholar's Hub for a quick Trending Papers overview.

Scoops of the last weeks

• GaussianPro: 3D Gaussian Splatting with Progressive Propagation
• Grounded SAM: Assembling Open-World Models for Diverse Visual Tasks
• Adaptive Ensembles of Fine-Tuned Transformers for LLM-Generated Text Detection
• Scaling Rectified Flow Transformers for High-Resolution Image Synthesis
• StarCoder 2 and The Stack v2: The Next Generation
• GaLore: Memory-Efficient LLM Training by Gradient Low-Rank Projection
• SPHINX-X: Scaling Data and Parameters for a Family of Multi-modal Large Language Models
• MathVerse: Does Your Multi-modal LLM Truly See the Diagrams in Visual Math Problems?
• ShortGPT: Layers in Large Language Models are More Redundant Than You Expect
• Towards General Computer Control: A Multimodal Agent for Red Dead Redemption II as a Case Study

• LGM: Large Multi-View Gaussian Model for High-Resolution 3D Content Creation
• Yi: Open Foundation Models by 01.AI
• Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data
• OLMo: Accelerating the Science of Language Models
• Dolma: an Open Corpus of Three Trillion Tokens for Language Model Pretraining Research
• Lumiere: A Space-Time Diffusion Model for Video Generation
• Sora: A Review on Background, Technology, Limitations, and Opportunities of Large Vision Models
• VideoCrafter2: Overcoming Data Limitations for High-Quality Video Diffusion Models
• SGS-SLAM: Semantic Gaussian Splatting For Neural Dense SLAM
• IM-3D: Iterative Multiview Diffusion and Reconstruction for High-Quality 3D Generation
• KTO: Model Alignment as Prospect Theoretic Optimization

Mamba Family:

• VMamba: Visual State Space Model
• Vision Mamba: Efficient Visual Representation Learning with Bidirectional State Space Model
• VM-UNet: Vision Mamba UNet for Medical Image Segmentation
• The Hidden Attention of Mamba Models
• Mamba: Linear-Time Sequence Modeling with Selective State Spaces

To Take a Look

• Retrieval-Augmented Generation for Large Language Models: A Survey
• Real-World Robot Applications of Foundation Models: A Review
• Physics-Guided, Physics-Informed, and Physics-Encoded Neural Networks in Scientific Computing
• Finite Scalar Quantization: VQ-VAE Made Simple
• Benchmarking Neural Network Training Algorithms

Current Controversal Debates

• Will the Transformer principle (2017) be replaced by one new efficient technique: Mamba, RMKV, Retnet, ... ?
• Mixture of Experts or Dense model ? Dense Model, Sparse Model, Soft Sparse Model, which one is the best and for what ?
• Trends for Aligning LLMs to human preferences : RLHF, DPO, Self-Rewarding DPO, Self-Play Fine-Tuning (SPIN), KTO, which Fine-Tuning will become the reference ?
• Toward Multimodal Abilities : Add Modalities & Instruct LLMs (e.g. LLaVa, OneLLM) or Start from Multimodal Fundation Model (Flamingo-style), which way will be the future of AI ?
• Is Deep Reinforcement Learning dead ?
• Have PINNs ever been alive ?