The "Will GPT automate all the jobs?" paper is out With participation from OpenAI, OpenResearch and @penn 🧵 1/9

LLMs such as ChatGPT and BLOOMZ claim that they are multilingual, but does this mean they can generate code-mixed data? Follow this 🧵 to find out. (1/N) Paper: arxiv.org/abs/2303.13592

Congrats! Eric Rosen

📢Our paper on exploration for deep RL has been accepted as an Oral ICML Conference! 🎉 Count-based exploration enjoys strong theoretical foundations, yet it has been overshadowed by prediction-error methods like RND. Our embarrassingly simple idea could scale pseudocounts ⬇️ #ICML2023

Doktora öğrencimiz Alper'in çalışması SİU2023'de en iyi bildiri ödülü aldı. Tebrikler Alper Ahmetoglu! Boğaziçi Üni. Bilgisayar Mühendisliği Bölümü

BİZ TÜRKİYE'YİZ BİZ ŞAMPİYONUZ! 🇹🇷 Tribünlerde ve ekran başında sizler, sahada biz! Birlikte savaştık ve kazandık!! Bu şampiyonluk güzel ülkemiz için... Avrupa'nın en büyüğü TÜRKİYE! 🐺❤️🏆

Bu şampiyonluk her şeye değdi. Bu kızlara çok büyük ödüller verilmeli.

Cumhuriyet’in 100. yılında büyük gurur. Muhteşemsiniz.

🇹🇷 Şampiyon, kürsünün zirvesinde. #FileninSultanları

#EbrarKarakurt

Happy to share that our paper “Neural Field Movement Primitives for Joint Modelling of Scenes and Motions” has been accepted to #IROS2023. Joint work with Yasemin Bekiroglu and Marc P. Deisenroth. Project Page: fzaero.github.io/NFMP/ Arxiv: arxiv.org/abs/2308.05040 (1/6)

Günün yorumu: Saruman Dede :D

Keşke insanlar Emre Ugur gibi başarılı hocalarımızın nerede yemek yedikleriyle ilgileneceklerine, üniversitelerdeki kısıtlı imkanlara rağmen nasıl yıllardır Dünya çapında araştırma yaptıklarıyla ilgilenseler!

RL in POMDPs is hard because you need memory. Remembering *everything* is expensive, and RNNs can only get you so far applied naively. New paper: 🎉 we introduce a theory-backed loss function that greatly improves RNN performance! 🧵 1/n

How do robots understand natural language? #IJCAI2024 survey paper on robotic language grounding We situated papers into a spectrum w/ two poles, grounding language to symbols and high-dimensional embeddings. We discussed tradeoffs, open problems & exciting future directions!

🚨 What is the best way to use foundation models in robotics? Our new work shows that combining LLMs & VLMs with ideas from formal methods leads to robots that can verifiably follow complex, open-ended instructions in the real world. 🌍 We evaluate on over 150 tasks🚀 🧵 (1/4)

🤖🌟New GTA just dropped! Check out our latest paper on zero-shot skill transfer for robots. We show how modular task-axis controllers + visual foundation models enable real-world manipulation — without training or demos. 🌐 iamlab-cmu.github.io/GTA/ 📄 arxiv.org/abs/2505.11680

Can robots learn data-efficient world models for object manipulation? Our new paper, which won "Best Paper" at the Conference on Robot Learning RINO Workshop, shows how robots learn object-centric world models with just a few minutes of interaction data. 🧵1/7

Wouldn't it be great if robots were more data efficient? Manipulation has gotten more data efficient, but mobile manipulation (MoMa) is lagging. Introducing LAMBDA (λ): long horizon benchmark of realistically sized datasets to push the limits of MoMa models. 🤖 #IROS2025 🧵1/N

Why does most of hierarchical RL stop at the first layer of hierarchy—i.e. skills? Because until now, it wasn't clear how to learn the state abstractions in a principled way! Our recent NeurIPS paper shows how to start with a set of skills and learn state abstractions. 🧵1/8