DeepSeek R1 moment has come for GUI agents: Rule-based Reinforcement Learning gives better results than SFT with 500x smaller datasets! Traditionally (by which I mean "in the last few months"), GUI agents have been trained with supervised fine-tuning (SFT). This meant,
RESEARCH
-

Data Scaling Effects in Reinforcement Learning from Human Feedback
By
–
Exploring Data Scaling Trends and Effects in Reinforcement Learning from Human Feedback Shen et al.: https://
arxiv.org/abs/2503.22230 #ArtificialIntelligence #DeepLearning #ReinforcementLearning -

Physical AI Era Begins With NVIDIA Omniverse Innovation
By
–
Has the Era of Physical AI Begun? I sat down with Rev Lebaredian, #NVIDIA’s VP of Omniverse & Simulation Technology, at @NVIDIAGTC to explore the future of Physical AI! In this week’s #CXOSpice, discover how Physical AI is transforming industries with innovations like
-
EU Digital Europe Programme Work Plan 2025-2027 Released
By
–
Work Programme 2025-2027 of the Digital Europe Programme (DIGITAL) | Shaping Europe’s digital future https://
digital-strategy.ec.europa.eu/en/library/wor
k-programme-2025-2027-digital-europe-programme-digital
…
#digitaleu #digital #europe @DigitalEU @elaniazito @ArturHabant @CurieuxExplorer @Shi4Tech @PawlowskiMario @JolaBurnett @enilev @BetaMoroney -
Verifying Facts and Sources in AI-Assisted Internal Research
By
–
Most of my use is for internal research. I always verify facts and sources for key data points. And depending on the project, I will continue the conversation to expand on the research. Sometimes that’s where the real magic happens.
-
Evidence provision for Hall-Presenti report and Turing AI discussions
By
–
Yes, although I was only providing evidence to Hall-Presenti report then (no formal role). There's also later council discussions around Turing and AI Roadmap.
-

CPPO Accelerates Group Relative Policy Optimization Training
By
–
CPPO: Accelerating the Training of Group Relative Policy Optimization-Based Reasoning Models
Paper: https://
arxiv.org/pdf/2503.22342
Code: https://
github.com/lzhxmu/CPPO -

CPPO Boosts GRPO Speed by 8x on Mathematical Reasoning
By
–
GRPO just got a speed boost! Xiamen University introduced Completion Pruning Policy Optimization (CPPO), which significantly reduces the number of gradient calculations and updates.
How fast? On GSM8K, it's 8.32× faster than GRPO, and on MATH, the speedup is 3.51×. -

Baidu’s ERNIE 4.5 defeats OpenAI’s GPT-4.5 at Chinese chess
By
–
Baidu's ERNIE 4.5 model just destroyed OpenAI's GPT-4.5 at Chinese chess ERNIE won all three matches and was even seen “taking it easy” during portions of the lopsided victories!
-

Google Gemini 2.5 Pro Experimental Launches Surpassing Human IQ
By
–
Google's new Gemini 2.5 Pro Experimental model has started rolling out to all users, including those on the free tier
— Rowan Cheung (@rowancheung) 31 mars 2025
2.5 Pro also scored a 130 on Mensa Norway’s IQ test, well surpassing other AI models and the average human's score of 100pic.twitter.com/Xh0iG4M65JGoogle's new Gemini 2.5 Pro Experimental model has started rolling out to all users, including those on the free tier 2.5 Pro also scored a 130 on Mensa Norway’s IQ test, well surpassing other AI models and the average human's score of 100
