BioNLP Workshop - Revision history

Dina: /* IMPORTANT DATES */

2026-04-20T21:45:58Z

IMPORTANT DATES

← Older revision		Revision as of 21:45, 20 April 2026
Line 10:		Line 10:
	* Notification of acceptance: May 4 (Monday), 2026		* Notification of acceptance: May 4 (Monday), 2026
	* Camera-ready paper due: May 12 (Tuesday), 2026		* Camera-ready paper due: May 12 (Tuesday), 2026
			* Pre-recorded video due (hard deadline): June 4, 2026
	* Workshop: July 3 - 4, 2026, Collocated with ACL 2026 in San Diego, CA		* Workshop: July 3 - 4, 2026, Collocated with ACL 2026 in San Diego, CA
	<br>		<br>

Dina: /* IMPORTANT DATES */

2026-04-17T21:01:51Z

IMPORTANT DATES

← Older revision		Revision as of 21:01, 17 April 2026
Line 7:		Line 7:
	(All submission deadlines are 11:59 p.m. UTC-12:00 “anywhere on Earth”)		(All submission deadlines are 11:59 p.m. UTC-12:00 “anywhere on Earth”)

	* Paper submission deadline: April 17 (Friday), 2026		* Paper submission deadline: <s>April 17 (Friday)</s> April 20, Monday, 2026
	* Notification of acceptance: May 4 (Monday), 2026		* Notification of acceptance: May 4 (Monday), 2026
	* Camera-ready paper due: May 12 (Tuesday), 2026		* Camera-ready paper due: May 12 (Tuesday), 2026

Dina: /* IMPORTANT DATES */

2026-02-27T19:59:45Z

IMPORTANT DATES

← Older revision		Revision as of 19:59, 27 February 2026
Line 10:		Line 10:
	* Notification of acceptance: May 4 (Monday), 2026		* Notification of acceptance: May 4 (Monday), 2026
	* Camera-ready paper due: May 12 (Tuesday), 2026		* Camera-ready paper due: May 12 (Tuesday), 2026
	* Workshop: July 3 OR 4, 2026, Collocated with ACL 2026 in San Diego, CA		* Workshop: July 3 - 4, 2026, Collocated with ACL 2026 in San Diego, CA
	<br>		<br>
	Please watch for the updates!		Please watch for the updates!

Dina at 19:58, 27 February 2026

2026-02-27T19:58:49Z

← Older revision		Revision as of 19:58, 27 February 2026
Line 85:		Line 85:

	==== ClinicalSkillQA====		==== ClinicalSkillQA====
	~~This~~ task ~~extends evaluation~~ to a ~~multimodal setting. Given an image~~ of ~~a medical student’s procedure, a question,~~ and ~~four answer options, the goal is~~ for ~~participants to train a model to generate~~ the ~~correct response~~. The dataset ~~will be~~ constructed from ~~~80~~ video clips of medical student clinical procedures, collected from ~~a partner medical school~~. ~~This task provides a unified framework~~ for ~~benchmarking, diagnosing~~, and ~~advancing~~ LLM ~~capabilities~~ for ~~both clinical decision support and medical training~~. ~~Evaluation will follow a multiple~~-~~choice QA setup with accuracy as~~ the ~~primary metric, with additional stratified analyses by skill type and modality~~.		ClinSkill QA formulates clinical skill understanding and continuous perception for clinical skill assessment as an ordering task: the MLLM is required to arrange shuffled key frames into a coherent sequence of clinical actions and to provide explanations for the resulting order. The dataset is constructed from video clips of medical student clinical procedures, collected from Zhongnan Hospital of Wuhan University and Cofun. For evaluation, we use Task Accuracy (exact ordering) and Pairwise Accuracy (the fraction of adjacent pairs correctly ordered) for the ordering results, and BertScore as well as an LLM-as-judge(G-Eval) for assessing the quality of the ordering explanations.
			Dataset access: After joining our Google Group (https://groups.google.com/g/clinskill-qa2026), participants can download the dataset via Google Drive: https://drive.google.com/file/d/1PdaBthI3OeDEbMlxUAxKqsKWw_t6hdJ2/view?usp=drive_link

			Additional information: https://whunextgen.github.io/ClinicalskillQA/



	===Workshop Organizers===		===Workshop Organizers===

Dina: /* MedGenVidQA */

2026-02-26T19:12:52Z

MedGenVidQA

← Older revision		Revision as of 19:12, 26 February 2026
Line 68:		Line 68:

	====MedGenVidQA====		====MedGenVidQA====
	The task focuses on grounding answers with reference attribution to mitigate generation of false statements by LLMs when answering biomedical questions. BioGen 2026 introduces generation of multimodal answers from textual and visual sources with citations, leveraging PubMed and HealthVidQA as multimodal sources. The test set is based on the information requests submitted by self-identified non-clinicians to the MedlinePlus service provided by the National Library of Medicine. The evaluation will leverage BioACE, an automated metric that strongly correlates with human evaluation on the BioGen 2024 textual dataset.		The task focuses on grounding answers with reference attribution to mitigate generation of false statements by LLMs when answering biomedical questions. BioGen 2026 introduces generation of multimodal answers from textual and visual sources with citations, leveraging PubMed and HealthVidQA as multimodal sources. The test set is based on the information requests submitted by self-identified non-clinicians to the MedlinePlus service provided by the National Library of Medicine. The evaluation will leverage BioACE, an automated metric that strongly correlates with human evaluation on the BioGen 2024 textual dataset.

			The CodaBench registration and submission portal for the MedGenVidQA shared task is now open.

			Participants can access the test dataset and submit their system runs for evaluation through the portal.

			Task A: Multimodal Retrieval (MMR) https://www.codabench.org/competitions/13989/

			Task B: Multimodal Answer Generation (MAG) https://www.codabench.org/competitions/14014/

			Task C: Visual Answer Localization (VAL) https://www.codabench.org/competitions/14015/

			Submission Deadline: March 31, 2026

			More details can be found on the shared task webpage: https://medgenvidqa.github.io/

	==== ClinicalSkillQA====		==== ClinicalSkillQA====

Dina at 19:10, 26 February 2026

2026-02-26T19:10:35Z

← Older revision		Revision as of 19:10, 26 February 2026
Line 67:		Line 67:
	Task Homepage: https://psydefdetect-shared-task.github.io/		Task Homepage: https://psydefdetect-shared-task.github.io/

	====~~BioGen~~====		====MedGenVidQA====
	The task focuses on grounding answers with reference attribution to mitigate generation of false statements by LLMs when answering biomedical questions. BioGen 2026 introduces generation of multimodal answers from textual and visual sources with citations, leveraging PubMed and HealthVidQA as multimodal sources. The test set is based on the information requests submitted by self-identified non-clinicians to the MedlinePlus service provided by the National Library of Medicine. The evaluation will leverage BioACE, an automated metric that strongly correlates with human evaluation on the BioGen 2024 textual dataset.		The task focuses on grounding answers with reference attribution to mitigate generation of false statements by LLMs when answering biomedical questions. BioGen 2026 introduces generation of multimodal answers from textual and visual sources with citations, leveraging PubMed and HealthVidQA as multimodal sources. The test set is based on the information requests submitted by self-identified non-clinicians to the MedlinePlus service provided by the National Library of Medicine. The evaluation will leverage BioACE, an automated metric that strongly correlates with human evaluation on the BioGen 2024 textual dataset.


	==== ClinicalSkillQA====		==== ClinicalSkillQA====

Dina at 19:08, 26 February 2026

2026-02-26T19:08:40Z

← Older revision		Revision as of 19:08, 26 February 2026
Line 70:		Line 70:
	The task focuses on grounding answers with reference attribution to mitigate generation of false statements by LLMs when answering biomedical questions. BioGen 2026 introduces generation of multimodal answers from textual and visual sources with citations, leveraging PubMed and HealthVidQA as multimodal sources. The test set is based on the information requests submitted by self-identified non-clinicians to the MedlinePlus service provided by the National Library of Medicine. The evaluation will leverage BioACE, an automated metric that strongly correlates with human evaluation on the BioGen 2024 textual dataset.		The task focuses on grounding answers with reference attribution to mitigate generation of false statements by LLMs when answering biomedical questions. BioGen 2026 introduces generation of multimodal answers from textual and visual sources with citations, leveraging PubMed and HealthVidQA as multimodal sources. The test set is based on the information requests submitted by self-identified non-clinicians to the MedlinePlus service provided by the National Library of Medicine. The evaluation will leverage BioACE, an automated metric that strongly correlates with human evaluation on the BioGen 2024 textual dataset.

	==== ~~Clinical Skill QA~~====		==== ClinicalSkillQA====
	This task extends evaluation to a multimodal setting. Given an image of a medical student’s procedure, a question, and four answer options, the goal is for participants to train a model to generate the correct response. The dataset will be constructed from ~80 video clips of medical student clinical procedures, collected from a partner medical school. This task provides a unified framework for benchmarking, diagnosing, and advancing LLM capabilities for both clinical decision support and medical training. Evaluation will follow a multiple-choice QA setup with accuracy as the primary metric, with additional stratified analyses by skill type and modality.		This task extends evaluation to a multimodal setting. Given an image of a medical student’s procedure, a question, and four answer options, the goal is for participants to train a model to generate the correct response. The dataset will be constructed from ~80 video clips of medical student clinical procedures, collected from a partner medical school. This task provides a unified framework for benchmarking, diagnosing, and advancing LLM capabilities for both clinical decision support and medical training. Evaluation will follow a multiple-choice QA setup with accuracy as the primary metric, with additional stratified analyses by skill type and modality.

Dina at 19:08, 26 February 2026

2026-02-26T19:08:06Z

← Older revision		Revision as of 19:08, 26 February 2026
Line 61:		Line 61:
	If you have any question, feel free to send an email to medexact-acl2026+owner@googlegroups.com.		If you have any question, feel free to send an email to medexact-acl2026+owner@googlegroups.com.

	====PsyDefDetect~~: Detecting Psychological Defense Mechanisms in Conversations~~====		====PsyDefDetect====

	This task focuses on classifying Seeker’s utterances in supportive conversations into specific Psychological Defense Levels based on the Defense Mechanism Rating Scales (DMRS) framework. The benchmark addresses the challenge of capturing subtle linguistic cues of deep-seated psychological mechanisms within highly informal and context-dependent emotional dialogues. This initiative supports research at the intersection of clinical psychology and NLP, aiming to operationalize complex psychological constructs for computational analysis. Participating systems will be ranked using Accuracy, Precision, Recall, and F1-score.		Detecting Psychological Defense Mechanisms in Conversations. This task focuses on classifying Seeker’s utterances in supportive conversations into specific Psychological Defense Levels based on the Defense Mechanism Rating Scales (DMRS) framework. The benchmark addresses the challenge of capturing subtle linguistic cues of deep-seated psychological mechanisms within highly informal and context-dependent emotional dialogues. This initiative supports research at the intersection of clinical psychology and NLP, aiming to operationalize complex psychological constructs for computational analysis. Participating systems will be ranked using Accuracy, Precision, Recall, and F1-score.

	Task Homepage: https://psydefdetect-shared-task.github.io/		Task Homepage: https://psydefdetect-shared-task.github.io/

Dina: /* SUBMISSION INSTRUCTIONS */

2026-02-26T19:07:18Z

SUBMISSION INSTRUCTIONS

← Older revision		Revision as of 19:07, 26 February 2026
Line 26:		Line 26:
	Submissions need to be anonymous.		Submissions need to be anonymous.

	Submission site for the workshop: ~~START system (link coming soon)~~		'''Submission site for the workshop: https://softconf.com/acl2026/bionlp2026'''

			'''Submission site for the Shared Tasks: https://softconf.com/acl2026/bionlp2026-st'''

	Please follow the ACL formatting guidelines: https://github.com/acl-org/acl-style-files		Please follow the ACL formatting guidelines: https://github.com/acl-org/acl-style-files

	Dual submission policy: papers may NOT be submitted to the BioNLP workshop if they are or will be concurrently submitted to another meeting or publication.		Dual submission policy: papers may NOT be submitted to the BioNLP workshop if they are or will be concurrently submitted to another meeting or publication.

	===WORKSHOP TOPIC AND CONTEXT===		===WORKSHOP TOPIC AND CONTEXT===

Dina at 22:09, 16 January 2026

2026-01-16T22:09:17Z

@@ Line 3: / Line 3: @@
 ===IMPORTANT DATES ===
-<!-- * <font color="#808080"> Paper submission deadline: March 20, 2025
+<b>Tentative</b>
-* Notification of acceptance: April 28, 2025
+<br>
-* Camera-ready paper due: May 25, 2025 -- No extensions due to ACL publication deadline.
+(All submission deadlines are 11:59 p.m. UTC-12:00 “anywhere on Earth”)
-* Pre-recorded video due (hard deadline): July 10, 2025. Please contact Underline if you need help </font> -->
-* <b> Workshop: Collocated with ACL 2026 that will be held in San Diego, CA, from July 2 - 7, 2026.  </b>  More information will be announced soon.
+* Paper submission deadline: April 17 (Friday), 2026
 ===WORKSHOP TOPIC AND CONTEXT===
@@ Line 47: / Line 70: @@
 ==== Clinical Skill QA====
 This task extends evaluation to a multimodal setting. Given an image of a medical student’s procedure, a question, and four answer options, the goal is for participants to train a model to generate the correct response. The dataset will be constructed from ~80 video clips of medical student clinical procedures, collected from a partner medical school. This task provides a unified framework for benchmarking, diagnosing, and advancing LLM capabilities for both clinical decision support and medical training. Evaluation will follow a multiple-choice QA setup with accuracy as the primary metric, with additional stratified analyses by skill type and modality.