Jon Rawski

Researcher

Stony Brook University

About

I am a PhD candidate in the Linguistics Department at Stony Brook University, advised by Jeffrey Heinz. I am also a Graduate Assistant in the Center for Neural Circuit Dynamics, and a Junior Research Fellow at the Institute for Advanced Computational Science.

My work concerns the mathematics of human language and learning. These cognitive feats emerge from humans’ unique neuronal structure and computing power, allowing linguistic insight to contribute to the broader cognitive sciences and artificial intelligence. I frequently use tensor algebra, formal language theory, grammatical inference, differentiable machine learning, and finite model theory to clarify the nature of the human mental capacity of language.

Interests

Computational/Mathematical Linguistics
Cognitive Science
Artificial Intelligence

Education

PhD in Linguistics, 2021 (expected)
Stony Brook University
MSci in Cognitive Science, 2016
Higher School of Economics
BA in Linguistics, 2013
University of Minnesota

Recent & Upcoming Presentations

See my CV for a full list of all my talks.

Overcoming Poverty of Stimulus with Structure and Parameters

We show how phonological typology computationally emerges from combinations of relativized representations and the simplest of a suite of online learning algorithms.

May 8, 2020 2:00 PM Workshop on Theoretical Phonology (WTPh)

Slides

Understanding Machine Learning with Language and Tensors

I analyze the functional expressivity of sequential neural machine learning using tensor algebra and simulations on formal languages.

Feb 7, 2020 2:00 PM Dept. of Mathematics and Statistics, Concordia University

Slides

What Can Formal Languages Tell Us About Cognition?

I discuss the scope and limits of linguistic generalization in human and nonhuman animals using formal language theory.

Feb 6, 2020 4:00 PM Dept. of Linguistics, Concordia University

Projects & Publications

See full list of publications.

Phonological Abstractness in the Mental Lexicon

We overview the notion of phonological abstractness, various types of evidence for it, and consequences for linguistics and psychology.

Eric Bakovic, Jeffrey Heinz, Jon Rawski

Typology Emerges from Simplicity in Representations and Learning

We derive the well-studied subregular classes of formal languages, which computationally characterize natural language typology, purely from the perspective of algorithmic learning problems.

Dakotah Lambert, Jon Rawski, Jeffrey Heinz

Review Draft

The Logical Nature of Phonology Across Speech and Sign

This article examines whether the computational properties of phonology hold across spoken and signed languages, using model theory and logical transductions.

Jon Rawski

Review Draft Invited Talk Slides Learning Slides

What can formal language theory do for animal cognition studies?

We comment on mathematical fallacies present in artificial grammar learning experiments and suggest how to integrate psycholinguistic and mathematical results.

Aniello De Santo, Jon Rawski

Official Pub

Probing RNN Encoder-Decoder Generalization of Subregular Functions using Reduplication

We analyze the expressivity of a variety of recurrent encoder-decoder networks, showing they are limited to learning subsequential functions, and connecting RNNs with attention mechanisms to a class of deterministic 2-way transducers.

Max Nelson, Hossep Dolatian, Jon Rawski, Brandon Prickett

PDF Slides Official Pub

Multi-Input Strictly Local Functions for Templatic Morphology

We provide an automata-theoretic characterization of templatic morphology, extending strict locality to consider n-ary functions.

Hossep Dolatian, Jon Rawski

PDF Slides Official Pub

Multi-Input Strictly Local Functions for Tonal Phonology

We provide an automata-theoretic characterization of tonal phonology, extending strict locality to consider n-ary functions.

Jon Rawski, Hossep Dolatian

PDF Poster Official Pub

History of Phonology: Learnability

This chapter examines the brief but vibrant history of learnability in phonology.

Jeffrey Heinz, Jon Rawski

PDF

Tensor Product Representations of Subregular Formal Languages

I provide a vector space characterization of the Star-Free and Locally Threshold testable classes of formal languages, over arbitrary data structures.

Jon Rawski

PDF IJCAI slides NECPhon slides

Learning with Partially Ordered Representations

We describe a partial order on the space of model-theoretic constraints and a learning algorithm for constraint inference.

Jane Chandlee, Remi Eyraud, Jeffrey Heinz, Adam Jardine, Jon Rawski

PDF Poster Slides DOI Official Pub

Finite-State Locality in Semitic Root-and-Pattern Morphology

We describe the finite-state nature of root-and-pattern morphology using Semitic as a case study, and discuss issues of finiteness vs. infinity, and template emergence.

Hossep Dolatian, Jon Rawski

PDF DOI

No Free Lunch in Linguistics or Machine Learning: Reply to Pater

We caution about confusing ignorance of biases with absence of biases in machine learning and linguistics, especially for neural networks.

Jon Rawski, Jeffrey Heinz

PDF Official Pub

Quantified Sentences as a Window into Prediction and Priming: An ERP Study

We used event related potentials (ERPs) to examine the processing of quantified sentences in an auditory/visual truth value judgment task, specifically to probe truth value and quantifier type influences on the N400 and ERP markers of quantifier complexity.

Aniello De DeSanto, Jon Rawski, John Drury, Amanda Yazdani

PDF CLS Slides SNL Poster