@danielle_maddix Profile picture

Danielle Maddix

@danielle_maddix

ML at AWS AI Labs

Similar User
Bernie Wang photo

@LoVVgE

Qinqing Zheng photo

@qqyuzu

Cédric Archambeau photo

@cedapprox

Huan Wang photo

@huan__wang

Xueyan Zou photo

@xyz2maureen

Giovanni Zappella photo

@610v4nn1__

Alex Robey photo

@AlexRobey23

xiyuanzh photo

@xiyuanzh

Aaron Lou photo

@aaron_lou

Sheng Zha photo

@szha_

Tanya Marwah photo

@__tm__157

Haibin training LLMs photo

@eric_haibin_lin

PascalIv photo

@Iv_Pascal

Stephan Rabanser photo

@steverab

Danielle Maddix Reposted

How can we transfer knowledge from large foundation models to a much smaller downstream model? In new ICML work, we show a simple yet conceptually significant modification to knowledge distillation works well with negligible overhead! arxiv.org/abs/2406.07337 1/9

Tweet Image 1

Danielle Maddix Reposted

Are all these time-series-specific model design necessary for deep forecasters / foundation forecasting models? In Chronos, we claim no novelty in time series modeling, but that's exactly the point. "Everything should be made as simple as possible, but no simpler." #TimeSeries

🚀 Excited to share that we released Chronos today. Chronos is a framework for building pretrained time series models based on language model architectures. Simple idea: quantize time series into tokens and feed them into 🤗 @huggingface models. 🧵

Tweet Image 1


Danielle Maddix Reposted

🚀 Excited to share that we released Chronos today. Chronos is a framework for building pretrained time series models based on language model architectures. Simple idea: quantize time series into tokens and feed them into 🤗 @huggingface models. 🧵

Tweet Image 1

Danielle Maddix Reposted

Excited to share a project that I had the incredible opportunity to contribute to during my last summer internship at AWS: Chronos 📈! Chronos is a framework for pretrained probabilistic time series models that demonstrates exceptional zero-shot performance.

Tweet Image 1

Danielle Maddix Reposted

Excited to share our latest work! Chronos is a zero-shot forecasting model that can generate accurate predictions for new time series not seen during training. 📜Paper: arxiv.org/abs/2403.07815 💻Code: github.com/amazon-science… 🤗Model weights: huggingface.co/collections/am…

Tweet Image 1

The paper decisions are out for our ICLR 2024 Workshop on AI4DifferentialEquations In Science! See ai4diffeqtnsinsci.github.io/papers for details! Congratulations to all authors! #ai4science @iclr_conf


Danielle Maddix Reposted

📢 Final #CFP! Submit your research to the #AI4DifferentialEquations in Science Workshop @ #ICLR by Feb 10th! Excited to announce our partnership with @Algorithms_MDPI for a special issue featuring selected papers from our workshop. Stay tuned at ai4diffeqtnsinsci.github.io

Tweet Image 1

Reminder for a final call for papers for our AI4DifferentialEquations In Science 2024 ICLR Workshop due Saturday, February 10th AoE! Call for Papers: openreview.net/group?id=ICLR.… @iclr_conf #ai4science

Tweet Image 1

Danielle Maddix Reposted

Appreciations to @danielle_maddix for a captivating @AIforGood talk, shedding light on physics-constrained #ML for scientific computing 💻🌐 In case you missed it, watch it now on: youtube.com/live/qflj9ZPL1…

Tweet Image 1

Danielle Maddix Reposted

Ensure you catch the talk tomorrow at 5 PM! See you online🎙️ @AIforGood @danielle_maddix #reminder

Our next @AIforGood talk in the "AI for Earth and Sustainability Science" series is just #oneweek away! 🌎 Join us online on Jan 31 at 5 PM for a talk with @danielle_maddix, moderated by @Reichstein_BGC 👥 Register here: aiforgood.itu.int/event/physics-…

Tweet Image 1


Danielle Maddix Reposted

Our next @AIforGood talk in the "AI for Earth and Sustainability Science" series is just #oneweek away! 🌎 Join us online on Jan 31 at 5 PM for a talk with @danielle_maddix, moderated by @Reichstein_BGC 👥 Register here: aiforgood.itu.int/event/physics-…

Tweet Image 1

Danielle Maddix Reposted

At ICLR, Amazon scientists are hosting their first workshop on AI4DifferentialEquations in Science. They invite all submissions on using machine learning to solve differential equations with applications in science and engineering: amzn.to/3Hiyb11 #ICLR2024 #AI4Science

Tweet Image 1

Danielle Maddix Reposted

Machine learning (ML) models could help solve difficult problems in physics, but their outputs sometimes violate basic physical laws. In two papers, Amazon researchers show how to impose physics-based constraints on ML models. #MachineLearning #SciML


Danielle Maddix Reposted

Our team has research intern positions available for this fall (or late summer) at @awscloud AI Labs in Bay Area and Berlin, in the field of AI4Science / SciML, and time series forecasting with applications in supply chain optimization and cloud-based systems. DM if interested.


Our paper "Guiding continuous operator learning through Physics-based boundary constraints" is on arxiv (lnkd.in/gRDE5qFt)! Try our code here: lnkd.in/gXe6gm8h. Happy to have collaborated with Nadim Saad, Gaurav Gupta and Shima Alizadeh on this work.


Loading...

Something went wrong.


Something went wrong.