Artificial Intelligence for Legal Assistance
(AILA 2021)
Artificial Intelligence for Legal Assistance (AILA) is a series of shared tasks aimed at developing datasets and methods for solving a variety of legal informatics problems. Started in 2019 AILA has so far focused on statute/precedent retrieval and Rhetorical role labelling. This year again AILA will consist of two different tasks. While we retain last year's Rhetorical Role Labelling task, we add a new Legal Documents Summarization task. AILA 2021 will be collocated with Forum for Information Retrieval Evaluation (FIRE 2021) and will be a completely online event.
The task is to semantically segment a legal case document. More formally, it is a sentence classification task, where each sentence has to be assigned one of the 7 predefined labels or "rhetorical roles".
Given a court judgement, the task is to generate a summary by selecting the most important content from the judgements.
A more detailed description of the tasks can be found on the description pages of task 1 and task 2.
Related literature
Task 1
Bhattacharya, et al. "Overview of the FIRE 2020 AILA Track: Artificial Intelligence for Legal Assistance" FIRE 2020
Bhattacharya, et al. "Identification of Rhetorical Roles of Sentences in Indian Legal Judgments" JURIX 2019
Savelka et.al. "Segmenting U.S. court decisions into functional and issue specific parts" JURIX 2018
Nejadghoii et.al., "A semi-supervised training method for semantic search of legal facts in Canadian immigration cases" JURIX 2017
Task 2
TBA
General Guidelines:
A team can participate in either or both the tasks and further in either of the subtasks in task 2.
Each team can have at most 4 participants and can submit up to 3 different runs for each task
Each team must also submit a detailed description of their algorithm(s) along with run submission.
Use of pretrained embeddings is allowed ONLY if they are pretrained on a publicly available dataset OR the pretrained embedding models themselves are publicly available. Participants are NOT allowed to use any external resources other than pre-trained embeddings.