lil-lab Goto Github PK

amr

Cornell AMR Semantic Parser (Artzi et al., EMNLP 2015)

atis

bandit-qa

Code for Simulating Bandit Learning from User Feedback for Extractive Question Answering.

blocks

Blocks World -- Simulator, Code, and Models (Misra et al. EMNLP 2017)

cb-analysis

cb2

An NLP research and data collection platform.

cerealbar

Cereal Bar is a two-player web game designed for studying language understanding agents in collaborative interactions. This repository contains code for the game, a webapp hosting the game, the agent implementation, and recorded interactions in the game. http://lil.nlp.cornell.edu/cerealbar/

cerealbar_generation

chalet

Cornell House Agent Learning Environment

ciff

Cornell Instruction Following Framework

clif_cb

Continual learning for instruction following in CerealBar.

cogen

Learning from Feedback with Coupled Comprehension and Generation

drif

Dynamic Robot Instruction Following

dynet_tutorials

Contains various short notebooks showing how to use DyNet. Created for CS 5740 at Cornell University.

geoquery-corpus

The GeoQuery corpus

gsmn

Code for RSS2018 paper on the Grounded Semantic Mapping Network

hlsm

hlsm-1

icca

kilogram

The KiloGram Tangrams dataset

kilogram-annotation-task

Task website for collecting tangram annotations from MTurk.

lilgym

lilGym RL benchmark

lilgym-baselines

lm-class

Materials for a language modeling class, broadly construed

navi

Code for Weakly Supervised Learning of Semantic Parsers for Mapping Instructions to Actions (Artzi and Zettlemoyer, TACL 2013)

navigation-corpus

Navigation data used for Chen and Mooney 2011 and Artzi and Zettlemoyer 2013 (including cleaned up oracle data)

nccg

Neural Shift Reduce Parser for CCG Semantic Parsing (Misra and Artzi, EMNLP 2016)

newsroom

Tools for downloading and analyzing summaries and evaluating summarization systems. https://summari.es/

nlvr

Cornell NLVR and NLVR2 are natural language grounding datasets. Each example shows a visual input and a sentence describing it, and is annotated with the truth-value of the sentence.

peer-eval