|
|

|
ESIP Community Members suggest you consider submitting to the following AGU Fall Meeting sessions. To avoid an overload of individual messages to the ESIP-All Mailing List, we will share sessions here in the Monday Update each week.
- IN031 – Near Real-Time/Low Latency Data for Earth Science and Space Weather Applications: Near real time/low latency data and new big data techniques applied to satellite, airborne, marine (including uninhabited aerial/marine systems-UxS), and surface sensors are transforming existing end-user applications and spawning new ones. These applications demonstrate the utility of timely data and advanced analyses in diverse Earth and space science disciplines including weather prediction, flood and river forecasting, earthquake hazards and tsunami forecasting, volcanic eruptions, natural and human-caused hazards, public health, agriculture, marine, early warning, and space weather applications. In addition to traditional and emerging computer analyses, the use of apps for smartphones and tablets presents an opportunity to improve and expand the timely usage of data products and services. This session seeks contributions that demonstrate the benefit of near real time/low latency scientific or social media data, discuss innovative real time analysis approaches including machine learning and big data strategies, decrease data delivery latency, or identify gaps in current capabilities.
- IN039 – Solving Training Data Bottleneck for Artificial Intelligence/Machine Learning in Earth Science: While there are successful applications of Artificial Intelligence/Machine Learning (AI/ML) in Earth Science, the wider adoption of AI/ML has been limited. The challenge is no longer the lack of algorithms, tools, or computing resources, but rather the dearth of training data. Access to training data for supervised learning is required to attract AI/ML practitioners to tackle Earth Science problems. Creating labeled data at sufficient scales to support AI/ML algorithms is still a bottleneck and new strategies to increase training data size and diversity need to be explored. This session seeks submissions from AI/ML practitioners and data curators using different approaches or existing products to create new datasets. This session will enable the practitioners to share successful approaches to scale the process of generating labeled datasets. We also seek submissions focusing on best practices for labeling and structuring data including catalog and standardization to benchmark and share training data.
|
|
Questions/comments? Reply directly to this note or click the button below to email us at staff@esipfed.org
ESIP is funded with support from NASA, NOAA, and the USGS.
|
|
|
|
|
|