{"id":2388,"date":"2021-06-29T13:20:39","date_gmt":"2021-06-29T21:20:39","guid":{"rendered":"https:\/\/live-cometml.pantheonsite.io\/blog\/comet-newsletter-7-tensorboardx-integration-tesla-cvpr-ml-job-interviews-multitask-learning\/"},"modified":"2021-06-29T13:20:39","modified_gmt":"2021-06-29T21:20:39","slug":"comet-newsletter-7-tensorboardx-integration-tesla-cvpr-ml-job-interviews-multitask-learning","status":"publish","type":"post","link":"https:\/\/www.comet.com\/site\/blog\/comet-newsletter-7-tensorboardx-integration-tesla-cvpr-ml-job-interviews-multitask-learning\/","title":{"rendered":"Issue 7: Comet + TensorboardX, Tesla at CVPR, Guide to ML Job Interviews"},"content":{"rendered":"\n

A new Comet integration with TensorBoardX, Tesla at CVPR, a new e-book on ML job interviews, more intelligent multi-task learning, and more<\/em><\/h3>\n

 <\/p>\n\n\n\n

Welcome to issue #7 of The Comet Newsletter!\u00a0<\/p>\n\n\n\n

This week, we share a new Comet integration with TensorBoardX<\/a> and take a closer look at Andrej Karpathy\u2019s CVPR talk on Tesla\u2019s vision-only approach to autonomous vehicles<\/a>.<\/p>\n\n\n\n

Additionally, we highlight Facebook\u2019s new open-source time series library<\/a>, an exceptional e-book on ML job interviews<\/a>, and a great piece on how to more intelligently do multi-task learning<\/a>.<\/p>\n\n\n\n

Like what you\u2019re reading? Subscribe here.<\/a><\/h3>\n\n\n\n

And be sure to follow us on Twitter<\/a> and LinkedIn<\/a> \u2014 drop us a note if you have something we should cover in an upcoming issue!<\/p>\n\n\n\n

Happy Reading,<\/p>\n\n\n\n

Austin<\/p>\n\n\n\n

Head of Community, Comet<\/p>\n\n\n

——————————–<\/center>\n\n\n

INDUSTRY<\/strong> | WHAT WE’RE READING | PROJECTS | OPINION<\/p>\n\n\n\n

From CVPR 2021: Tesla\u2019s Andrej Karpathy on a Vision-Based Approach to Autonomous Vehicles<\/a><\/h2>\n\n\n\n

 <\/p>\n\n\n\n

\n
\"\"\n
Image Source<\/a><\/figcaption>\n<\/figure>\n<\/div>\n\n\n\n

 <\/p>\n\n\n\n

Tesla is doubling down on its vision-first approach to self-driving cars and will stop using radar sensors altogether in future releases.<\/p>\n\n\n\n

From its inception, Tesla has taken a different approach to most other companies developing a self-driving car that doesn\u2019t rely on Lidar. Instead they used a combination of radar sensors and 8 cameras placed around the vehicle.<\/p>\n\n\n\n

In this talk at CVPR 2021, Andrej Karpathy explains the challenges with this approach and why Tesla has decided to focus on cameras<\/a> and stop using radar sensors. The main challenge with using different sensor types is sensor fusion. As Elon Musk puts it:<\/a> \u201cWhen radar and vision disagree, which one do you believe? Vision has much more precision, so better to double down on vision than do sensor fusion.\u201d<\/p>\n\n\n\n

Radar is used to track objects that are in front of the car and determine their distance, speed, and acceleration. The approach taken by the Tesla team has been to compute this information from onboard cameras rather than radar sensors, and they claim to have developed a new vision-based system that outperforms existing radar-based approaches.<\/p>\n\n\n\n

In order to achieve this, they relied on both auto-labelling techniques and the power of their fleet. Broadly, the approach is as follows:<\/p>\n\n\n\n

    \n
  1. Use all the information at your disposal (radar, cameras, hindsight, etc) to create a neural network that will generate ground-truth labels<\/li>\n
  2. Use these labels to train a neural network that can be deployed on cars (strict latency requirements, can’t use data from the future, etc.)<\/li>\n
  3. Run this neural network in shadow mode on the Tesla fleet<\/li>\n
  4. Based on 200 handcrafted triggers, identify edge-cases for which the predictions might be incorrect<\/li>\n
  5. Annotate edge cases and go back to step 1<\/li>\n<\/ol>\n\n\n\n

    Even with this complex vision-based approach, moving away from radar sensors is a risky move. It will be fascinating to see how this plays out over the next couple of years.<\/p>\n\n\n\n

    For a more complete look at the most interesting updates from CVPR 2021, check out this recap in Analytics India Magazine<\/em><\/a>.<\/p>\n