Two evaluation tasks

TemporalWebLab Tasks

TemporalWebLab includes two tasks: estimating how soon an answer may need updating, and answering questions about past information and changes using Web archives. Teams may participate in either task independently or in both.

01 Classification · Ordinal prediction

Question Recency Requirement Estimation

Estimate how soon the currently correct answer to a question may need to be updated using one of twelve ordered recency classes.

Input
A question, with accompanying context where provided
Output
One of 12 ordered recency classes
Evaluation
Macro-F1
View Task 1 details
02 Retrieval · Evidence-grounded QA

Question Answering over Web Archives

Answer questions about information at specific times and changes across archived versions of Web pages using evidence from the provided archive corpus.

Input
A question and access to the provided archive corpus
Output
An answer and supporting archive record IDs
Evaluation
Answer correctness and evidence grounding
View Task 2 details