The WGFAST meeting is fast approaching!
Thank you to everyone who completed the initial registration questionnaire. We now invite all participants to submit an abstract for either a poster or an oral presentation, and to sign up for the workshops taking place ahead of the main event.
Abstract submission deadline: 6 March
Participants who registered for the mentor/mentee program will receive a separate email with further details soon.
Please note that the WGFAST main meeting runs Monday 20th of April through Friday 24th of April 2026, with Friday reserved for WGFAST business.
Workshop Schedule
Saturday, 18 April
9:00-12:00 - echopype Hackathon (Part 1) - Wu-Jung
13:00-17:00 - AI Workshop (Part 1) - Benoit
Sunday, 19 April
9:00-12:00 -echoSMS Datastore and Model Selection - Gavin & Mike
13:00-14:00 - GOATS - Ben
14:00-16:00 - Discussion on Data Processing Pipelines/Workflows and Data Levels - Wu-Jung & Beth
16:00-18:00 - AI Workshop (Part 2) - Benoit
Accommodation and Dinner
Information about recommended accommodation options and the conference dinner will be shared as soon as it becomes available.
We’re looking forward to seeing you all there!
Workshop Details
echopype Workshop
Bring your own data and tell us about your needs. This hands-on workshop aims to demonstrate what is currently possible within the echopype environment, identify community needs, and explore how you can contribute to further development.
AI Workshop – Background
Large amounts of split-beam EK60/EK80 CW data are collected annually and are also available historically. This has fostered the development of frameworks for efficient data pre-processing and AI-based data processing (segmentation and classification) across the WGFAST community. Several pipelines are now well developed and already used at scale.
This workshop will provide hands-on demonstrations using selected datasets to share experience with different pipelines, improve accessibility, and encourage interoperability. A longer-term goal is to evaluate AI-based processing against classical approaches under different conditions to better understand the strengths and limitations of AI methods.
Workshop objectives:
-
Align with standards (e.g., data format, annotations)
-
Share experience with frameworks for data preparation, pre-processing, and AI data processing
-
Compare segmentation and classification performance across frameworks using a benchmarking dataset (AI methods as the focus, with deterministic approaches as baseline)
Pipelines considered:
-
Echopype
-
LSSS/CRIMAC
-
Ifremer
Proposed timeline:
-
Feb–Mar (online, core group): framework presentations and workshop planning
-
WGFAST 2026 (2 half-days)
-
-
Open session: overview, scope, and community discussion
-
Core group: hands-on hackathon session
-
echoSMS – Datastore and Model Selection
Explore the echoSMS package for target strength modelling, with special focus on the Datastore and model selection. This workshop follows up on previous echoSMS workshops and hackathons. Bring your own shapes, questions, and code.
GOATS
This session continues discussions from the GOATS workshop in 2025, working toward a comprehensive global assessment of Target Strength (TS) data for fish and invertebrates. The goal is to review and synthesize existing research to improve understanding, accessibility, and application across diverse user groups.
Discussion on Data Processing Pipelines, Workflows, and Data Levels
This session will focus on currently operable processing pipelines, workflows, and data levels. Interoperability, shareability, and transparency would benefit from clearer community definitions of pipelines and data levels. Let’s discuss how we can move toward more standardized and interoperable definitions. See here for a previously proposed level definition.
