Mads Rydahl

I’m excited to announce the launch of Florent Venture Partners and that I am joining as Founding LP, Chief Science Officer, and IC member. @CDorffer @LarsRas Bloomberg has the full story:

Try Proem News: Trustworthy perspectives producthunt.com/posts/proem-ne… by @_proem

Attending a founder event at #techbbq with Adam Cheyer.

@AziziShekoofeh @GoogleDeepMind @GoogleAI @Google @GoogleHealth @taotu831 @vivnat @JMZambranoC proem.ai/paper/arxiv/24…

Excited to share Tx-LLM, a Large Language Model for Therapeutics. Developing therapeutics is a long road, and we're excited about the possibility of AI to contribute. arxiv.org/abs/2406.06316

@KKawaharazuka proem.ai/paper/arxiv/24…

Our paper presented at Robotics and Automation Magazine is now on arXiv and YouTube! Toward Autonomous Driving by Musculoskeletal Humanoids: A Study of Developed Hardware and Learning-Based Software arxiv: arxiv.org/abs/2406.05573 youtube: youtube.com/watch?v=qQqv2p…

@arankomatsuzaki proem.ai/paper/arxiv/24…

ShiftAddLLM: Accelerating Pretrained LLMs via Post-Training Multiplication-Less Reparameterization 80% memory and energy reductions over the original LLMs arxiv.org/abs/2406.05981

@arankomatsuzaki proem.ai/paper/arxiv/24…

Autoregressive Model Beats Diffusion: Llama for Scalable Image Generation Achieves 2.18 FID on ImageNet 256×256 benchmarks, outperforming the popular diffusion models such as LDM, DiT proj: github.com/FoundationVisi… abs: arxiv.org/abs/2406.06525

@_reachsumit proem.ai/paper/arxiv/24…

Should We Fine-Tune or RAG? Evaluating Different Techniques to Adapt LLMs for Dialogue Evaluates techniques like in-context learning, fine-tuning, and knowledge grounding for adapting LLMs to different dialogue types, finding no universal best approach 📝arxiv.org/abs/2406.06399

@_reachsumit proem.ai/paper/arxiv/24…

UMBRELA: UMbrela is the (Open-Source Reproduction of the) Bing RELevance Assessor Presents an open-source toolkit that demonstrates the ability of LLMs to perform relevance assessment tasks with human-quality judgments. 📝arxiv.org/abs/2406.06519 👨🏽‍💻github.com/castorini/umbr…

@osanseviero proem.ai/paper/arxiv/24…

RecurrentGemma 9B by Google is out 🔥 ⚡️Super fast for long sequences: Good throughput+latency 👀Base and instruct tuned versions 🏆Similar quality as Gemma Check the y-axis below 🤯 Models: huggingface.co/collections/go… Griffin paper: arxiv.org/abs/2402.19427