Helm.ai is building the next generation of AI technology for automation.

Honda Motor

Binnacle Partners

A.Capital Ventures

Amplo

One Way Ventures

Kevin Durant

Lee Linden

Navneet Dalal

Nicolas Berggruen

Vlad Tenev

Mando Corporation

Freeman Group

Base Capital Funding

Mark Leslie

Neil King

ACVC Partners

Sungwoo Hitech

Goodyear Ventures

Helm.ai

Helm.ai will be exhibiting at CES next week at booth # 4265 in West Hall.

Come by for a 10 min live presentation or email ces2024@helm.ai to schedule a private demo!

Along with our foundation models for intent and path prediction for urban driving (https://lnkd.in/g3PaF8Xs), another new technology we'll be showcasing is our generative AI approach to simulation for semantic segmentation and perception.

A core challenge in building cutting edge perception DNNs for automotive production requirements is simultaneously getting high accuracy across a wide variety of image appearances, including different times of day, weather conditions, different object appearances/geographical settings/levels of traffic and many other variations.

In the demo below (play in 4K for full resolution), we show our ability to transform real images into highly realistic simulated versions of those images with desired appearances across relevant variations:

Scenario Transfer: Highway -> Urban -> Highway
Time of Day Transfer: Day -> Night
Weather Transfer: Average Weather -> Rain, Fog and Snow
Vehicle In-painting: Higher Traffic -> Lower Traffic

Importantly, the segmentation masks accurately fit the original and transformed images, which means we are able to generate highly realistic synthetic labeled data that can be used for large scale DNN training and validation. 

#helmai #ces2024 #autonomousdriving #selfdrivingcars #robotics #generativeai #computervision #artificialintelligence

https://lnkd.in/g8pvwtai


We're excited to announce our generative AI based foundation models for intent prediction and path planning for self driving cars. 

At Helm.ai we're pioneering a scalable AI approach that seamlessly transitions from mass production for ADAS L2/L3 all the way to large scale L4 deployments within the same framework.   

Perception is the critical first component of any self-driving stack. The more comprehensive, accurate and temporally stable a perception system is, the easier it is to build the downstream prediction capabilities, which is especially critical for driving in complex urban environments. 

Leveraging our industry-validated surround-view urban perception system and Deep Teaching training technology, we trained DNN foundation models for intent prediction and path planning to learn directly from real driving data, allowing them to understand a wide variety of urban driving scenarios and the subtleties of human behavior without the need for traditional physics based simulators or hand-coded rules.

See the press release below for more information and a video demo. 

#generativeai #selfdrivingcars #autonomousdriving  #artificialintelligence #computervision 

https://lnkd.in/g3PaF8Xs

We have a few brand new developments we will be showcasing at our #ces2024 booth. Be sure to stop by. You won’t want to miss it! West Hall #4265

Perception is the critical first component of any self-driving stack. Moreover, the more robust and temporally stable your perception system is, the easier it is to build downstream parts of the stack for intent prediction and path planning. 

Below is a demonstration of our surround view perception system deployed in urban environments, including 3D bounding box detection for vehicles and full scene semantic segmentation of freespace, vehicles, pedestrians, road markings and lanes with semantic classification, crosswalks, curbs, sidewalks, traffic lights, traffic signs, buildings, and vegetation. 

This system was built using our Deep Teaching foundation models for semantic segmentation and 3D understanding, which leverage large scale unsupervised training. 

While average case highway driving is relatively simple, urban driving presents a much bigger challenge due to the vast increase in complexity of road rules, different types of objects, behaviors of other vehicles and pedestrians, and road geometries. In particular, the number of perception features increases by an order of magnitude and the ability to predict what other drivers and pedestrians will do next and to plan the optimal path becomes a much harder problem in city settings. 

At Helm.ai we're building an AI-first approach to high end ADAS software that scales all the way to mass L4 deployments. Stay tuned for upcoming announcements of our new Deep Teaching DNN foundation models for intent prediction and path planning for tackling autonomous driving in urban environments. 

#helmai #autonomousdriving #ai #computervision #artificialintelligence #selfdrivingcars #adas 

Stop by our booth in Vegas to see some advanced demos and learn more about what's under the hood at Helm.ai. See you there! Booth #4265 West Hall #ces2024 #autonomousvehicles #av #ai #machinelearning #softwareinnovation 

Helm.ai is pioneering a breakthrough in unsupervised learning for AI and autonomous technologies. Their approach to driverless car AI has garnered attention and funding, positioning them as a key player in the autonomous driving industry.

AI:HELM

Helm.ai

Filings Not Available