We took the challenge of reading one paper a day for two weeks and here is the list of the papers we all read. For the sake of exactness, it was around 1 paper per day on average and we didn’t need to read the same papers, some people shared some papers but in general, we were fully free to choose.

After the 2 weeks, we made short meetings to share a very short summary of each of the papers

Here is the list of all papers discussed

  1. DROID-SLAM

  2. GitHub - ISEE-Technology/CamVox: [ICRA2021] A low-cost SLAM system based on camera and Livox lidar.

  3. A Review of Visual-LiDAR Fusion based Simultaneous Localization and Mapping

  4. Masked Autoencoders are Scalable Vision Learners

  5. Blind Geometric Distortion Correction on Images Through Deep Learning

  6. Self-supervised Monocular Depth Estimation with internal feature fusion

  7. Recurrent Multi-View alignment network for unsupervised surface registration

  8. Prototypical cross-attention network for multiple object tracking and segmentation

  9. Nerf

  10. In-place scene labelling and understanding with implicit scene representation

  11. Nerf in the wild

  12. Recurrent Multiframe single shot detector for video object detection

  13. LoFTR: Detector-Free Local Feature Matching with Transformers

  14. Skip-Convolutions for Efficient Video Processing

  15. One-Shot Free-View Neural Talking-Head Synthesis for Video Conferencing

  16. Probabilistic Future Prediction for Video Scene Understanding

  17. Urban Driving with Conditional Imitation Learning

  18. FIERY

  19. Spatial Transformer Networks

  20. ADOP

  21. Variational End-to-End Navigation and Localization

  22. Neural Point-Based Graphics

  23. ORB-SLAM2

  24. ORB-SLAM: a Versatile and Accurate Monocular SLAM System

  25. Plen-octrees

  26. Lift, Splat, Shoot: Encoding Images from Arbitrary Camera Rigs by Implicitly Unprojecting to 3D

  27. MonoLayout: Amodal scene layout from a single image

  28. Orthographic Feature Transform for Monocular 3D Object Detection

  29. The Transformer Model in Equations

  30. Every Model Learned by Gradient Descent Is Approximately a Kernel Machine

  31. MotionNet: Joint Perception and Motion Prediction for Autonomous Driving Based on Bird’s Eye View Maps

  32. Why Having 10,000 Parameters in Your Camera Model is Better Than Twelve

  33. ViViT: A Video Vision Transformer

  34. Semantic-assisted 3D Normal Distributions Transform for scan registration in environments with limited structure

  35. Discrete Kalman Filter Tutorial

  36. Relational inductive biases, deep learning, and graph networks (in progress)

  37. Inductive Biases for Deep Learning of Higher-Level Cognition.

  38. On the Measure of Intelligence

  39. Unsupervised Learning of Visual 3D Keypoints for Control

  40. Transfer learning based few-shot classification using optimal transport mapping from preprocessed latent space of backbone neural network

  41. Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks

  42. Prototypical Networks for Few-shot Learning