Become a sponsor to AIStream
AIStream aims to to democratize access to deep learning techniques as well as apply them to high impact areas. At the moment many deep learning techniques are simply published at research conferences but then never utilized or only leveraged by large companies that can afford full-time research labs 😞.
In particular, right now AIStream is focused on developing open source deep learning for 📈time series forecasting, classification, and anomaly detection frameworks/systems. We hope that these tools can help small businesses, non-profits, and researchers in disciplines with temporal data (ecology, healthcare, hydrology, etc) leverage the full power of deep learning models for time series analysis and prediction. Secondly, we are also focused on applying these tools to a number of high impact AI4Good areas. Specifically we are interested in forecasting COVID-19 spread and the impact of policy interventions🦠, predicting flash floods and droughts 🏞️, and predicting patient vitals/risk of decline in the ICU 🤒.
All of AIStream's work is open, transparent and well documented. Sponsorship is important as it helps pay for cloud infrastructure for experiments, living expenses of core contributors (some of whom dedicate tremendous amounts time), and funding sprints/other developer events.
With five sponsors we should be able to cover the majority of our cloud computing costs. These cloud computing costs allow us to rigorously test frameworks like Flow-Forecast to make sure they work in a distributed setting. They also allow us to continue to benchmark our time series models on more real world time series datasets so that you can truly know, which models work the best.
Meet the team
-
Isaac Godfried isaacmgI love contributing to open source projects and would work full-time on them if I could
Featured work
-
AIStream-Peelout/flow-forecast
Deep learning PyTorch library for time series forecasting, classification, and anomaly detection (originally for flood forecasting).
Python 2,059 -
AIStream-Peelout/Water
Water is an attempt to create an open-access dataset and ML framework for monitoring/forecasting stream flows and snowpack depth.
Jupyter Notebook 7
0% towards 5 monthly sponsors goal
Be the first to sponsor this goal!
$5 a month
SelectMicro-creek: Five dollars per month is enough to cover basic storage costs of data in GCP. For this level of contribution we will include your name in our list of supporters.
$30 a month
SelectCreek: Thirty dollars per month is enough money to pay for the cost of a GPU to run a model hyper-parameter search on GCP (at least for relatively small searches). It is important to benchmark models supported by Flow-Forecast on real world datasets so that users can know how well they work in comparison with one another. For this level of contribution we will include your name in our list of supporters.