dRISK presents breakthrough performance in retraining autonomous vehicles to detect high risk "edge cases" 13.04.2021: dRISK emerges from stealth, presents breakthrough performance in retraining autonomous vehicles to detect high risk "edge cases" dRISK, a London and Pasadena-based startup

dRISK emerges from stealth, presents breakthrough performance in retraining autonomous vehicles to detect high risk "edge cases"

dRISK, a London and Pasadena-based startup that has so far been in stealth, today announced its launch and that it has - for the first time commercially - employed its edge case retraining tool to achieve a 6x performance in time to detect high-risk events for autonomous vehicles (AVs).



Whereas semi-autonomous and autonomous vehicles currently do not always detect high-risk events in time to react to them (oncoming cars peeking into the lane from behind other vehicles, vehicles running red lights concealed by other cars), dRISK's tools for retraining AVs to recognize edge cases represent a dramatic step forward in the ability to retrain autonomous vehicles to well outperform humans at even the trickiest driving scenarios. The results were formally presented at NVIDIA's GTC conference on April 12, 2021.



dRISK's mission is to help make AVs dramatically safer as soon as possible. The new patented knowledge graph technology (analagous to Google's knowledge graph of the internet, but in dRISK's case a knowledge graph of real-world events) solves a number of problems which have plagued AV developers so far -- encoding massively high-dimensional data from all the different relevant data sources into a tractable form, and then offering the full spectrum of edge cases so as to retrain on not just with what has already happened but will happen in the future.



dRISK delivers simulated and real+simulated edge cases in semi-randomized, impossible-to-game training and test sequences, to achieve superior testing and retraining results for customers on real-life data.



Unlike traditional training and development techniques, in which AVs are trained to recognize primarily whole entire vehicles and pedestrians under advantageous lighting conditions, with dRISK's edge cases AVs are trained to recognize just the predictors of high-risk events (e.g. the headlights of an oncoming car peeking into the lane amid low-visibility).



AV systems trained this way can recognize high-risk events sooner, without a significant decrease in performance on low-risk events.



dRISK's customers include AV developers, major transport authorities and one of the world's largest insurers, all of whom have an interest in mitigating AV risk and improving AV performance. dRISK has so far delivered semi-customized solutions on an individualized basis, but intends to release a version of its AV retraining product directly on the web later this spring to open up these capabilities to the wider AV community.

dRISK Inc and wholly-owned subsidiary dRISK.ai Limited, along with its partners in the UK-based D-RISK consortium, Imperial College London's Transport Systems Laboratory, Claytex and DG Cities, received £3M in funding from the UK's Centre for Connected and Autonomous Vehicles to develop the ultimate driver's test for the self-driving car.



In addition to its government funding, dRISK raised seed funding in a closed round led by Okapi Ventures, with Netsu Equity Ltd, Poetic Partners, SaaS Ventures and Mount Wilson Ventures. dRISK holds 4 patents on its technology, with two additional patents pending. dRISK's CEO is Chess Stetson, Ph.D. (Computation and Neural Systems, Caltech) and its team boasts talent from Stanford, Berkeley, Harvard, Oxford, Cambridge, Imperial College London and NASA.




See more here: dRISK’s “Risk Aware Perception”




dRISK also photos