understand.ai

03/02/2022

How to label 20+ million objects in 12 weeks’ time at 98% quality? Is it possible to overcome the bottleneck of manual work?

5 years ago UAI was founded to solve the problem of manual annotations with the help of . We predicted the exponential rise in data volume needed for the evolution of . We knew that it'd need a mature AI-technology to reach the required throughput and annotation quality.

✔️We are happy to confirm that our UAI Automation Engine passed the real-world ADAS validation project test.

Combined with the extensive project experience, we are able to give automotive OEMs and Tier-1s the edge in the autonomous driving race. 🏁

Discover how 23 million objects labeled in 12 weeks at 98% quality changed what’s possible in data annotation for autonomous driving. A case study.

11/01/2022

High-volume image has just become 2022 ready 🍾

We're proudly presenting the new release of the ✨UAI Anonymizer✨, our neural network powered image anonymization tool. Developed to handle the exponential growth of data volume in , the new version is:

🚀🚀🚀🚀 4x faster
😷 has a +99% identification rate for faces with and without face masks
☁️ runs on all the big 3
📸 is much more versatile - camera perspectives, lens geometry, and resolution
🔄 has upgraded data exchange methods for faster & more flexible data processing

See for yourself and get a demo today https://hubs.la/Q011NTYK0

23/12/2021

We wish you a happy Holiday Season and a successful 2022 🎉 We'll keep on doing everything in our powers to make a reality. For autonomous trucks to deliver your gifts 🎁 and for ADAS to ease your drive home for holidays 🏡. Proud to be a part of the movement and thankful for all who helped us shape our business. #2022

13/12/2021

L3 are here. More autonomy, higher safety requirements, more thorough validation. It all means that much, much more data will be needed develop and test every autonomous driving function.

Last week Mercedes-Benz broke the news about a Level 3 autonomous driving system launch in Germany at the beginning of 2022. Stellantis announcement of their plans to include Level 3 autonomous driving capability in theirs car starting in 2024 followed. The tech is developed in partnership with BMW.

The shift of liability from driver to manufacturer seen in the transition from L2 to L3 marks a tipping point in terms of validation effort required. Companies developing ADAS/ AD will have to tackle increasingly larger volumes of data to validate their products. This is not a question of more time or more resources to process the data. It’s a threshold of what you are able to do and what is impossible with traditional approach. This is the point in time when extensive annotation automation is needed to succeed.

08/12/2021

What happens when machine learning model enters the real world?

>>> Tweaking your model has limitations once for the use in real world applications, because it’s often not good enough. To make it really helpful, you'll have to focus on the data. It's becoming more and more important to maintain your dataset and make it representative for real world problems. The real world variety changes the classical paradigm of software development. It’s called Software 2.0. >>>

Read Daniels list of best practices we apply to make our AI production ready and discover how we approach data quality, quantity and distribution.
Go to our blog https://hubs.la/Q010tM7X0

We’ve built a production-ready AI-based automation engine for data annotation. How do we get our models into production?

24/11/2021

An amazing team of , and experts is looking for new members. The list of open positions is long:

*Senior Project Manager for Funded Projects >>> for research aficionados
*Customer Success Managers + Inside Sales + Head of Sales >>> for true customer whisperers
*Machine Learning, Full-stack, Back-end, DevOps, Site Reliability, Customer Success and Cloud engineers >>> for code benders & architects
*HR Business Partner >>> for people empowerers

For everyone who'd like to make AI accessible for real world applications - we are
Apply today https://hubs.la/H0_M0CC0

17/11/2021

understand.ai goes Las Vegas! Meet uns at CES 2020 at the dSPACE booth no. 9263. Discover the newest progress in automating for . Celebrate the comeback of in-person human contact!!! and book yourself a meeting slot at https://hubs.la/H0_bBcz0

15/11/2021

1.2 million 3D object annotations labeled at 99% quality in one month. Learn how UAI helped Mercedes-Benz AG to make a training and validation project a success story.

1.2 mm 3D objects labeled at 99% quality in a month. Learn how UAI helped Mercedes-Benz AG  to turn a training & validation project into a success story.

04/11/2021

Why is an accurate representation of reality important in simulation? Let Thorsten take you on a roundtrip through the Scenario Generation service by UAI and dSPACE. Discover how to leverage your test & validation activities with real-world data.

The real world is complex. Transferring this complexity into your simulation and test systems is key for the successful validation of ADAS/AD functions.In ou...

02/11/2021

What are the 3 critical points that determine the success of a large scale annotation projects according to L&T Technology Services Limited? How is UAI along with LTTS shaping the future of and AI-powered data annotation?

Check out https://hubs.la/H0-s3z00

19/10/2021

has recently risen in popularity thanks to the adoption by various and Deep Learning approaches. So why do ground truth annotations matter?

Read https://hubs.la/H0ZDfkL0

14/10/2021

Before you do anything with your recorded data, data protection regulations such as GDPR, APPI and CCPA require you to remove critical personal information from street scene data collected during driving.
You'll need a state-of-the-art tool to automate it. Our AI-powered tool depersonalizes recorded image data by detecting and blurring faces and license plates in a wide range of situations. The UAI Anonymizer is a SaaS solution powered by multiple deep learning algorithms enabling high image throughput and detection rates. The solution is flexible in its use, customizable to cover edge cases, and fully integrated into the UAI data annotation pipeline.

Designed to support large scale data collection for . compliance made in Germany.

See for yourself. Demo request link in the comments ↓