Repository for MuSiQue: Multi-hop Questions via Single-hop Question Composition

Last update: Jan 2, 2023

Overview

🎵 MuSiQue: Multi-hop Questions via Single-hop Question Composition

This is the repository for our paper "MuSiQue: Multi-hop Questions via Single-hop Question Composition". We'll release the more data and code here.

Installation

conda create -n musique python=3.8 -y && conda activate musique

Data

You can run the following script or download it manually from here.

bash download_data.sh

The result will be stored in data/ directory. It contains train and dev sets of MuSiQue-Ans and MuSiQue-Full and their sample predictions files.

Evaluation

You can use evaluate_v0.1.py to evaluate your predictions against ground-truths. For eg.:

python evaluate_v0.1.py data/musique_ans_v0.1_dev_sample_prediction.jsonl data/musique_ans_v0.1_dev.jsonl

Citation

If you use this in your work, please cite use:

@article{trivedi2021musique,
  title={MuSiQue: Multi-hop Questions via Single-hop Question Composition},
  author={Trivedi, Harsh and Balasubramanian, Niranjan and Khot, Tushar and Sabharwal, Ashish},
  journal={arXiv preprint arXiv:2108.00573},
  year={2021}
}

You might also like...

Repository for "Improving evidential deep learning via multi-task learning," published in AAAI2022

Improving evidential deep learning via multi task learning It is a repository of AAAI2022 paper, “Improving evidential deep learning via multi-task le

11 Nov 19, 2022

A library for answering questions using data you cannot see

A library for computing on data you do not own and cannot see PySyft is a Python library for secure and private Deep Learning. PySyft decouples privat

8.5k Jan 2, 2023

The official implementation for ACL 2021 "Challenges in Information Seeking QA: Unanswerable Questions and Paragraph Retrieval".

Code for "Challenges in Information Seeking QA: Unanswerable Questions and Paragraph Retrieval" (ACL 2021, Long) This is the repository for baseline m

25 Oct 30, 2022

Comments

codes about model

Hello！I am very glad to see this work.The dataset has highly quality and question decompose methods are very effective. I also study multi-hop reading comprehension and my postgraduate's projects is solve multi-hop reading comprehension via decompose multi-hop question. When I read your work's code.I can not find code about model.There are only have the evalution step code. Could you please release your work code about model ? Thanks a lot.

opened by fanruiwen 2

Repository for MuSiQue: Multi-hop Questions via Single-hop Question Composition

Related tags

Overview

🎵 MuSiQue: Multi-hop Questions via Single-hop Question Composition

Installation

Data

Evaluation

Citation

You might also like...

Repository for "Improving evidential deep learning via multi-task learning," published in AAAI2022

A library for answering questions using data you cannot see

The official implementation for ACL 2021 "Challenges in Information Seeking QA: Unanswerable Questions and Paragraph Retrieval".

A pytorch implementation of Reading Wikipedia to Answer Open-Domain Questions.

Python Interview Questions

EMNLP'2021: Simple Entity-centric Questions Challenge Dense Retrievers

EMNLP'2021: Simple Entity-centric Questions Challenge Dense Retrievers

Solutions and questions for AoC2021. Merry christmas!

Code for "Single-view robot pose and joint angle estimation via render & compare", CVPR 2021 (Oral).

Comments

codes about model

Owner

The dataset and source code for our paper: "Did You Ask a Good Question? A Cross-Domain Question IntentionClassification Benchmark for Text-to-SQL"

EMNLP 2021: Single-dataset Experts for Multi-dataset Question-Answering

EMNLP 2021: Single-dataset Experts for Multi-dataset Question-Answering

AugMax: Adversarial Composition of Random Augmentations for Robust Training

Image-Stitching - Panorama composition using SIFT Features and a custom implementaion of RANSAC algorithm

ConE: Cone Embeddings for Multi-Hop Reasoning over Knowledge Graphs

Reinforcement-learning - Repository of the class assignment questions for the course on reinforcement learning

(CVPR 2022 - oral) Multi-View Depth Estimation by Fusing Single-View Depth Probability with Multi-View Geometry

Official repository with code and data accompanying the NAACL 2021 paper "Hurdles to Progress in Long-form Question Answering" (https://arxiv.org/abs/2103.06332).

Code of U2Fusion: a unified unsupervised image fusion network for multiple image fusion tasks, including multi-modal, multi-exposure and multi-focus image fusion.