Quantifying social biases in NLP: A generalization and empirical comparison of extrinsic fairness metrics

P Czarnowska, Y Vyas, K Shah - Transactions of the Association for …, 2021‏ - direct.mit.edu
Measuring bias is key for better understanding and addressing unfairness in NLP/ML
models. This is often done via fairness metrics, which quantify the differences in a model's …

Anti-efficient encoding in emergent communication

R Chaabouni, E Kharitonov… - Advances in Neural …, 2019‏ - proceedings.neurips.cc
Despite renewed interest in emergent language simulations with neural networks, little is
known about the basic properties of the induced code, and how they compare to human …

A survey on emergent language

J Peters, CW de Puiseau, H Tercan… - arxiv preprint arxiv …, 2024‏ - arxiv.org
The field of emergent language represents a novel area of research within the domain of
artificial intelligence, particularly within the context of multi-agent reinforcement learning …

Toward More Human-Like AI Communication: A Review of Emergent Communication Research

N Brandizzi - IEEE Access, 2023‏ - ieeexplore.ieee.org
In the recent shift towards human-centric AI, the need for machines to accurately use natural
language has become increasingly important. While a common approach to achieve this is …

EGG: a toolkit for research on Emergence of lanGuage in Games

E Kharitonov, R Chaabouni, D Bouchacourt… - arxiv preprint arxiv …, 2019‏ - arxiv.org
There is renewed interest in simulating language emergence among deep neural agents
that communicate to jointly solve a task, spurred by the practical aim to develop language …

Capacity, bandwidth, and compositionality in emergent language learning

C Resnick, A Gupta, J Foerster, AM Dai… - arxiv preprint arxiv …, 2019‏ - arxiv.org
Many recent works have discussed the propensity, or lack thereof, for emergent languages
to exhibit properties of natural languages. A favorite in the literature is learning …

What they do when in doubt: a study of inductive biases in seq2seq learners

E Kharitonov, R Chaabouni - arxiv preprint arxiv:2006.14953, 2020‏ - arxiv.org
Sequence-to-sequence (seq2seq) learners are widely used, but we still have only limited
knowledge about what inductive biases shape the way they generalize. We address that by …

Co-evolution of language and agents in referential games

G Dagan, D Hupkes, E Bruni - arxiv preprint arxiv:2001.03361, 2020‏ - arxiv.org
Referential games offer a grounded learning environment for neural agents which accounts
for the fact that language is functionally used to communicate. However, they do not take into …

The curious case of representational alignment: Unravelling visio-linguistic tasks in emergent communication

T Kouwenhoven, M Peeperkorn, B Van Dijk… - arxiv preprint arxiv …, 2024‏ - arxiv.org
Natural language has the universal properties of being compositional and grounded in
reality. The emergence of linguistic properties is often investigated through simulations of …

Nellcom-x: A comprehensive neural-agent framework to simulate language learning and group communication

Y Lian, T Verhoef, A Bisazza - arxiv preprint arxiv:2407.13999, 2024‏ - arxiv.org
Recent advances in computational linguistics include simulating the emergence of human-
like languages with interacting neural network agents, starting from sets of random symbols …