“LazImpa”: Lazy and Impatient neural agents learn to communicate efficiently

Previous work has shown that artificial neural agents naturally develop surprisingly non-efficient codes. This is illustrated by the fact that in a referential game involving a speaker and a listener neural networks optimizing accurate transmission over a discrete channel, the emergent messages fail to achieve an optimal length. Furthermore, frequent messages tend to be longer than infrequent ones, a pattern contrary to the Zipf Law of Abbreviation (ZLA) observed in all natural languages. Here, we show that near-optimal and ZLA-compatible messages can emerge, but only if both the speaker and the listener are modified. We hence introduce a new communication system, “LazImpa”, where the speaker is made increasingly lazy, i.e., avoids long messages, and the listener impatient, i.e., seeks to guess the intended content as soon as possible.

[1]  George Kingsley Zipf,et al.  Human behavior and the principle of least effort , 1949 .

[2]  Eugene Kharitonov,et al.  Compositionality and Generalization In Emergent Languages , 2020, ACL.

[3]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[4]  Steven T Piantadosi,et al.  Word lengths are optimized for efficient communication , 2011, Proceedings of the National Academy of Sciences.

[5]  Eugene Kharitonov,et al.  Anti-efficient encoding in emergent communication , 2019, NeurIPS.

[6]  Ray Jackendoff,et al.  A Parallel Architecture perspective on language processing , 2007, Brain Research.

[7]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[8]  G. Zipf.  The Psycho-Biology Of Language: AN INTRODUCTION TO DYNAMIC PHILOLOGY , 1999 .

[9]  E. Gibson,et al.  How Efficiency Shapes Human Language , 2019, Trends in Cognitive Sciences.

[10]  Jing Peng,et al.  Function Optimization using Connectionist Reinforcement Learning Algorithms , 1991 .

[11]  Thomas M. Cover,et al.  Elements of information theory (2. ed.) , 2006 .

[12]  Yingying Wen,et al.  A compression based algorithm for Chinese word segmentation , 2000, CL.

[13]  Alexander Peysakhovich,et al.  Multi-Agent Cooperation and the Emergence of (Natural) Language , 2016, ICLR.

[14]  David Lusseau,et al.  Compression as a Universal Principle of Animal Behavior , 2013, Cogn. Sci..

[15]  W. Marslen-Wilson.  Functional parallelism in spoken word-recognition , 1987, Cognition.

[16]  José M. F. Moura,et al.  Natural Language Does Not Emerge ‘Naturally’ in Multi-Agent Dialog , 2017, EMNLP.

[17]  Karl J. Friston.  The free-energy principle: a unified brain theory? , 2010, Nature Reviews Neuroscience.

[18]  Michael L. Anderson,et al.  The problem with brain GUTs: conflation of different senses of "prediction" threatens metaphysical disaster. , 2013, The Behavioral and brain sciences.

[19]  J. Weijer,et al.  Word length, sentence length and frequency: Zipf revisited , 2004 .

[20]  Jelena Mirkovic,et al.  Incrementality and Prediction in Human Sentence Processing , 2009, Cogn. Sci..

[21]  Pieter Abbeel,et al.  Gradient Estimation Using Stochastic Computation Graphs , 2015, NIPS.

[22]  Gabriel Altmann,et al.  Word Length and Word Frequency , 2007 .

[23]  Marco Baroni,et al.  How agents see things: On visual representations in an emergent language game , 2018, EMNLP.

[24]  Jessica R McLachlan,et al.  Speedy revelations: how alarm calls can convey rapid, reliable information about urgent danger , 2020, Proceedings of the Royal Society B.

[25]  Stephen Clark,et al.  Emergence of Linguistic Communication from Referential Games with Symbolic and Pixel Input , 2018, ICLR.

[26]  Simon Kirby,et al.  Spontaneous evolution of linguistic structure-an iterated learning model of the emergence of regularity and irregularity , 2001, IEEE Trans. Evol. Comput..

[27]  Ronald J. Williams,et al.  Simple Statistical Gradient-Following Algorithms for Connectionist Reinforcement Learning , 2004, Machine Learning.

[28]  Eugene Kharitonov,et al.  EGG: a toolkit for research on Emergence of lanGuage in Games , 2019, EMNLP.

[29]  Kara D. Federmeier.  Thinking ahead: the role and roots of prediction in language comprehension. , 2007, Psychophysiology.

[30]  Simon Kirby,et al.  Zipf’s Law of Abbreviation and the Principle of Least Effort: Language users optimise a miniature lexicon for efficient communication , 2017, Cognition.

[31]  Ivan Titov,et al.  Emergence of Language with Multi-agent Games: Learning to Communicate with Sequences of Symbols , 2017, NIPS.

[32]  Territoire Urbain,et al.  Convention , 1955, Hidden Nature.

[33]  A. Clark.  Whatever next? Predictive brains, situated agents, and the future of cognitive science. , 2013, The Behavioral and brain sciences.