This week we are presenting our work on timestamp supervision for action recognition at IROS and ECCV L2ID 2022. Check out the following three minute introductory video. PDF: arxiv.org/abs/2206.15031 YouTube: youtube.com/watch?v=cohbno…

RetroActivity: The Industry 4.0 Comic In this second episode, a moving automotive assembly line in Detroit, MI, is encountering several expensive quality challenges. Ally and Jack work together to identify a scalable solution for the problems. retrocausal.ai/retroactivity2

In this third episode, a factory in Cleveland, OH where medical devices get assembled and packed in kits, is facing severe quality and takt time challenges. Download the full episode: retrocausal.ai/retroactivity3 #industry40 #smartfactory #manufacturing #computervision

In this fourth episode of RetroActivity, a Contract Manufacturer in Shenzhen assembles electronics devices for OEMs, and needs to maintain quality, while frequently re-deploying their lines. Mei is the smart innovation engineer who has a solution! retrocausal.ai/retroactivity4

Transform your #manufacturing process & boost #productivity, #quality, & #safety with #standardwork using our cutting-edge Pathfinder platform. Features incl. flexible def, operator training, deviation alerts, prescriptive analytics & more. 15-30% increase in productivity! #lean

Ever wanted to feel like a #factoryhero? Now's your chance! Join us at the Annaheim Convention Center to learn how to assemble a network switch with ease. Pathfinder will guide you every step of the way and catch any mistakes you make - no prior experience required! #MDMWest2023

Missed us at MD&M West last week? No worries! Here's a 2-minute video showing our demo. We've added "projected work instructions" and IoT tooling to our Pathfinder platform and demonstrated these for the first time to general public at the conference. youtu.be/z5gT-j0dgk4

Check out our dope new intro vid filmed in the Seattle showroom! It's fire 🔥🎬 Don't miss it! #Seattle #showroom #newvideo #dope youtu.be/HN3bXHeEmTY

📣 Introducing Kaizen Copilot! 🚀 The Generative AI agent for Continuous Process Improvement in Manufacturing 🏭 Check out our Web and Demo: retrocausal.ai/kaizen/ #GenerativeAI #industry40

Did you miss us at Automate 2023 last week? Fear not! Andrey Konin, our Chief Architect and Co-Founder walks you through a part of our demo there! youtube.com/watch?v=GYa4Xk…

A major challenge for unsupervised learning of procedural activities is due to steps that are performed out-of-order, missing steps, and step repetitions. We're releasing our paper which directly addresses these challenges. PDF: arxiv.org/abs/2305.19478 youtube.com/watch?v=ZgvbwD…

We are excited to share this latest work on video-based self-supervised learning which performs 2D skeleton-based temporal video alignment and shows superior performance on various fine-grained activity understanding tasks. PDF: arxiv.org/pdf/2305.19480… youtu.be/HQ1i-M0KiSU

We're absolutely thrilled to announce our new $5.3M financing round to meet the growing demand for our manufacturing process optimization platform. Press Release: retrocausal.ai/funding-announ…

Missed MD&M West this year? No worries! Here's a quick look at our booth at the leading conference for Medical device manufacturers. #mdmwest2024 #medicaldevices #scrapcosts #yields #workertraining youtube.com/watch?v=CiCD79…

Boost operator efficiency with Assembly Copilot! Assembly Copilot uses computer vision to guide assembly without disrupting workflow. youtube.com/shorts/zvZKnk9…

Thrilled to announce that Retrocausal has been selected as one of the top four manufacturing technology companies featured in CB Insights' prestigious AI 100 list for 2024! #AI100 #ManufacturingTech #Innovation

We got to learn from the achievements of operators and IEs who presented their Ergonomics-focused Kaizens at the Ergo Cup 2024 finals last month! Here are our picks from the line up, together with Copilot's assessments of their respective improvements. t.ly/tcVcR

Our work on self-supervised machine learning for building models of diverse human activities in a privacy-preserving manner has passed peer-review and will be presented at the European Conference on Computer Vision #ECCV2026 2024 conference in Milan, Italy. youtu.be/k0ysYs6Yl6M

Watching a long video by peeking at a few frames, it’s easy to miss details and lose track of events. That’s the core issue video models face—leading to inaccurate predictions. Our latest work addresses this problem! arXiv: arxiv.org/abs/2412.02930 YT: youtu.be/lEUluMdNHcc

Thank you Siemens USA for your wonderful partnership, and for this great video that sums up why we do what we do so perfectly 🤝