Matthias Heller
@Matt_Heller_Co-Founder @LightlyAI | Harvard, HSG, HEC Paris | Posting about technology and business
Similar User
@ISusmelj
@LightlyAI
@LongTonyLian
@olesia_andersen
@SomilMathur14
@Mayank60963632
@devAtulSingh
@LittleMsBrainy
@big_berny
@Ali_Naseem2
🚀 In our latest article, we're looking at video data to catch up on recent developments in Self-Supervised Pre-Training. We cover the fundamental methods such as VideoMAE and its extensions. 🔗Link: lightly.ai/post/self-supe…
🔍 Read about the recent trends in Knowledge Distillation using Self-Supervised Learning. We see significant performance gains at reduced costs by distilling knowledge from large models into smaller, efficient ones. 🔗 Read more here: lightly.ai/post/knowledge…
🚀 We’ve revamped our documentation! Check it out if you want to learn how to load data using various dataset formats. We cover Torchvision datasets, @huggingface hub datasets, and custom datasets. 👉 Read more here: docs.lightly.ai/self-supervise…
The wall of developer ❤️ is live. Your messages motivate us to keep pushing the envelope. bloop.ai
We are excited to sponsor @usenix @SREcon APAC 2023 here in Singapore! Drop by the booth to jam on all things incident management or ask us about our private reliability leaders dinner we are hosting!
The #AI4Code reading group is back next week with @noahshinn024 the author of the Reflexion: A novel reinforcement learning framework. Reflexion converts binary/scalar feedback into verbal textual summaries, to be used as additional context for future LLM agent executions 👇
Check out @bloopdotai - LLM-based approaches to code search are much more useful than traditional and will be more widespread
We love building our docs, but sometimes wish detailed answers could be found more easily in our sprawl... We were lucky enough to get a sneakpeak of @trybloop at @RustNationUK and its pretty neat!
We’ve just reached the milestone of ⭐1,000 Github stars⭐ on our project! github.com/evidence-dev/e… To mark the occasion, we’re contributing $1,000 to support open source software
Great book from @vishakhhegde about 3D Deep Learning with Python
With everything happening with Twitter, not sure how long this message is going to last 😅. Happy to share a book I co-authored: 3D Deep Learning with Python. Amazon link: amzn.to/3Eo7iXy. Do the right thing and share this with your friends and colleagues.
We added support for the Masked Siamese Network (MSN) to @lightly MSN learns image representations by comparing cluster assignments of masked and unmasked image views. Great work by @mcaron31 @p_bojanowski @imisra_ @armandjoulin docs.lightly.ai/examples/msn.h… github.com/lightly-ai/lig…
A friend just shared this with me. A great example for how deep learning models struggle with out-of-distribution data. I wonder how many carriages the model has seen during training 😅
We are incredibly impressed by the great video with which we may conclude @ESADownstreamG1 series "New Space Stories". We'd like to thank Giulia Bordacchini and ESA Commercialisation Gateway for the great cooperation and ESA BIC Switzerland for working with us.
It seemed unrealistic to analyse, select, filter and enrich massive amounts of #data from #satellites and #EarthObservation. @LightlyAI has been working to become the data toolbox for #machinelearning, assisting #artificialintelligence companies. lightly.ai
How Lightly is making AI and Machine Learning work for your business I had the pleasure to interview for Entrepreneur Media's podcast. It was an absolute blast! You can find the episode here: entrepreneur.com/listen/action-… Check it out!
I’ll be at @CVPR in New Orleans next week. Exciting to see what progress has been made in computer vision over the last year and how the whole field is moving towards more data centric AI. Who else is going to CVPR? What workshops do you recommend? DM if you want to meet!
Almost got emotional when I received that swag from @ycombinator today! (The t-shirt we received when we started our batch in 2018 was « build something people want »). What a ride since then
Recently, I have been thinking a lot about data pipelines in computer vision and how to automate them. This article is part of a series, where I share my definition of a data pipeline, describe how to build one, and provide reasons for their importance. lightly.ai/post/what-is-a…
Find out more about how Lightly helped predict precipitation with improved accuracy in Part 2 of the blog series in collaboration with @Meteomatics! lightly.ai/post/predictin…
Read about how @Meteomatics used @LightlyAI to help predict rain with satellite images in Part 1 of a new blog series, by our ML engineer, Philipp Wirth. Stay tuned for Part 2 coming next week! Special thanks to @Meteomatics for the great collaboration! lightly.ai/post/predictin…
8 out of the 9 top stories on NYTimes are negative right now. Google News even worse…. I know there are serious problems in the world but I wonder if it’s healthy for the billion+ people online to read largely negative news about the world every day…
United States Trends
- 1. #GivingTuesday 19 B posts
- 2. Good Tuesday 30,8 B posts
- 3. #tuesdayvibe 3.212 posts
- 4. #alieninvasion N/A
- 5. Elvis 17,3 B posts
- 6. #TuesdayFeeling 1.023 posts
- 7. Delaware 73,1 B posts
- 8. #3Dic 2.330 posts
- 9. Jameis 63,2 B posts
- 10. Daniel Penny 83,9 B posts
- 11. Starfire 1.875 posts
- 12. Manfred 4.702 posts
- 13. Lemmy 1.237 posts
- 14. Watson 21,2 B posts
- 15. Golden At-Bat 1.069 posts
- 16. Vermont 16,4 B posts
- 17. Roddy 4.533 posts
- 18. Judges 72,2 B posts
- 19. Jeudy 31,8 B posts
- 20. Bovaer 83,9 B posts
Something went wrong.
Something went wrong.