75 mph Tesla crash occurred while Autopilot

 

Keep Thinking!!!

Self-driving fridge, hit by self-driving Waymo, recorded by self-driving Tesla

 

 

 Keep Going!!!

Humanoids

#humanoids = #llm powered decisions + #robotics + custom models trained for specific tasks, Build custom models to suit specific needs https://t.co/sxcaqCk46U

— Siva (@sivaram2k10) May 12, 2024

 

Keep Exploring!!!

Tesla CVPR Talk - [CVPR'22 WAD] Keynote - Ashok Elluswamy, Tesla

Key Notes

• Tesla autopilot for lane management, safety, lane changes, exits, blind spots
• 100k cars have self-driving beta software (parking lots, city streets, highway)

• Algos onboard computer
• 8 Cameras - 36 frames / second

• 360 views around cars
• No other sensors outside this
• Segmentation to classify Drivable / Not

• Depth is also not enough / Difficult for occlusion detection
• (Capture)
• Occupancy Networks

• Consistent output space
• Compute efficiently

• 3D point occupied or not
• Fixed queries attend over image space feature
• Positional embedding in image space

• Moving / Non-Moving obstacle
• Produce both moving/static in same frame
  
  
  
  
• Occupancy produces both include in occupancy (Static + Dynamic)

• Nerf - Scene creation
• Extract descriptions from scene
• Build perspective outside the world

• Occupancy to Collision Avoidance
• Based on direction and velocity estimate collision avoidance

A Survey of Deep Learning Techniques for Autonomous Driving

Twelve Questions about Tesla Autopilot Technologies Exposed at 2021 AI Day

Keep Exploring!!!

Interesting JD - DataScience Roles - Evaluation Software Engineer

I have come across several d#atascience #jobs, This one Evaluation Software Engineer is very interesting

The JD States below things

• Organize neural network challenge scenarios and route them to the appropriate evaluation suites
• Collaborate with engineers and program managers to identify which neural network challenge cases are the highest priority to improve
• Investigate if the challenge persists in newer versions of models

My version of understanding

For a vision model for a failed use case - pedestrian not detected, vehicle not detected they triage / prioritize / address

• Why a use case fails, How to enrich the dataset
• Do the key regions are activated when we interpret feature activation across layers
• Prioritize / Add data / customize network if needed / train / validate it is fixed

This reflects how much every scenario is validated, prioritized, and ensured models reflect the real-world scenarios. Most of the time we see ML, DL jobs but not this level of details and clarity.

The JD link

This is the difference between prototype vs production vs updates and how forward-looking they are in the future to handle all scenarios :), Behind all #autopilot models there would be tons of #scenarios and multiple Evaluation Software Engineers and automated suites validating it.

I have never seen a similar type of JD anywhere except Tesla :)

Keep Exploring!!