Andrej Karpathy (Tesla): CVPR 2021 (see description for better quality version)
Channel:YarrowFlower · Length:37:46 · Published:2021-06-21 · Views:113772
Summary
Andrej Karpathy's CVPR 2021 presentation on Tesla's Autopilot, covering its vision-based approach, sensor choices and the scale of Tesla's deep learning compute cluster. This upload is a lower-quality copy of the talk.
Why HN recommends it
HN cites it as the primary source for Tesla's vision-only reasoning, LIDAR arguments and sensor fusion challenges.
it's quite long, but the whole thing is quite interesting — HN, 2022
Self-driving & autonomous vehicles · Computer vision, speech & NLP
Subtitles (excerpt)
i can see andre is the screen sharing working i hope so let's try i'm embarrassed to say that i barely use zoom i know this is shocking yes [Music] okay yeah seems still to work even with limited experience okay so uh it's a pleasure to introduce andre uh i have to say i think he doesn't really need a lot of introduction uh but yeah um he is the senior director of ai at tesla and he basically supervises uh and leads the team responsible for all the neural networks uh in the autopilot so with this uh it's a pleasure to have you here and i'm really looking forward to your talk uh yeah great thank you for the introduction and i'm very happy that i was able to display some impressive zero shot learning on zoom okay uh yes i'm very excited to be here thank you for the introduction um i would like to talk a bit more about what tesla has been up for the last few months um so first to start uh i guess we are here at the cbpr uh workshop on autonomous driving so i think i'm preaching to the core a little bit i'd like to start with some slides on like why we're doing all of this and i think unfortunately we're kind of like in a bad shape when it comes to transportation in society and uh basically the issue is that we have these metallic
On Hacker News
I would really like to hear Tesla to defend their decision to not use them. Andrej Karpathy talks some about it in this (it's quite long, but the whole thing is quite interesting): [VIDEO] — caf, 2022-02
IIRC there was a presentation from Karpathy talking about the challenges with sensor fusion, particular in resolving divergence between e.g. the vision and the radar stack: [VIDEO] My background is in physics, but I find myself having a growing appreciate for the vision-only stack. It's really challenging building a formal understanding of the world that is robust to outliers that are so numerous… — agoose77, 2022-07
Stereo vision is oversold in humans. You can easily drive with one eye. We use a lot of information to estimate depth including relative size of things, as well as “structure from motion” which is the relative movement of things over time. Tesla does estimate depth using structure from motion (and probably other cues). They call it virtual LIDAR in presentations. [VIDEO] — TaylorAlexander, 2022-03
First off, amazing farm-bot project! I am looking forward to reading the details on your site. Thx for the pointers on Tesla. Had not seen the Hydranet stuff. There was a Karpathy talk about 2 weeks back at a CVPR workshop .. he revealed the scale of Tesla's current generation deep learning cluster [1]. It is insane! Despite being in industrial research, I don't foresee ever being able to touch a… — brutus1213, 2021-07
[VIDEO] In the above video Andrej Karpathy claims that LIDAR required previously compiled maps, the creation of which is not scalable. Do you disagree with that point of view? — adolph, 2021-07
Related videos
- Movie Accent Expert Breaks Down 32 Actors' Accents | WIRED · WIRED · 16:41
- How Self-Driving Cars will Destroy Cities (and what to do about it) · Not Just Bikes · 54:37
- Bay Area Vision Meeting: Unsupervised Feature Learning and Deep Learning · Google TechTalks · 48:10
- Google Self-Driving Car Project | SXSW Interactive 2016 · SXSW · 52:21
- "Voice Driven Development: Who needs a keyboard anyway?" by Emily Shea · Strange Loop Conference · 41:17
- Andrej Karpathy - AI for Full-Self Driving at Tesla · Matroid · 30:08
- Tensorflow and deep learning - without a PhD by Martin Görner · Devoxx · 2:35:53
- JPEG DCT, Discrete Cosine Transform (JPEG Pt2)- Computerphile · Computerphile · 15:11
- Fourier Optics used for Optical Pattern Recognition · Huygens Optics · 18:30
- George Hotz: Comma.ai, OpenPilot, and Autonomous Vehicles | Lex Fridman Podcast #31 · Lex Fridman · 1:59:36
- Traffic Has a Perfect Solution. Humans Are the Problem. · CGP Grey · 4:30
- Let's build the GPT Tokenizer · Andrej Karpathy · 2:13:34
- Massive-scale online collaboration | Luis von Ahn · TED · 16:40
- Breaking The Creepy AI in Police Cameras · Benn Jordan · 35:49
- !!Con 2017: HDR Photography in Microsoft Excel?! by Kevin Chen · Confreaks · 11:46
- Dasher: information-efficient text entry · Google TechTalks · 53:45