Research & Publications

Academic and applied research across machine learning, continual learning, and computational biology. Includes published work, active preprints, and working papers.

Publications

Peer-reviewed publications in academic journals.

Communications Biology2025
Published

A long non-coding RNA Leat1 mediates the hormone responsiveness of EfnB2 during male urogenital development

Deidre Mattiske, Pascal Bernard, Paul E. Gradie, Richard R. Behringer, Paul A. Overbeek, Rachel J. O'Neill, Tiffany Phillips, Melanie Stewart, Neil Youngson, Gerard Tarulli, Andrew J. Pask

We identify Leat1 as a conserved long non-coding RNA that regulates urethral development through its interaction with EfnB2. Testosterone upregulates Leat1 expression while estrogen suppresses it, and loss-of-function of Leat1 results in a severe hypospadias phenotype. These findings identify Leat1 as a novel molecular regulator of urethral closure and a potential target through which endocrine disruption may cause congenital urogenital anomalies.

Sexual Development2015
Published

A Comprehensive Atlas of the Adult Mouse Penis

Tiffany Phillips, Danielle Wright, Paul E. Gradie, Lindsey Johnston, Andrew J. Pask

A comprehensive anatomical and histological reference atlas of the adult mouse penis, providing systematic characterisation of penile tissue types, structural layers, vascular architecture, and innervation patterns. This atlas serves as a reference standard for studies of urogenital development, reproductive biology, and the long-term consequences of developmental perturbations on adult penile morphology.

Phytologia2009
Published

Morus murrayana (Moraceae): A New Mulberry from Eastern North America

S.J. Galla, B.L. Viers, P. Gradie, L. de Saar

We describe Morus murrayana, a new species of mulberry (Moraceae) from eastern North America. Morphological characterisation distinguishes M. murrayana from related taxa in the genus Morus through a combination of leaf morphology, fruit characters, and geographic distribution. A formal taxonomic description, synonymy, and notes on ecology and distribution are provided.

Patents

Granted patents from applied AI and machine learning research.

United States Patent Office2022
Patent

Semi-Supervised, Deep-Learning Approach for Removing Irrelevant Sentences from Text in a Customer-Support System

Anh Thien Dinh, Wei Chu, Paul E. Gradie

A system and method for processing customer support requests by segmenting text into sentences, encoding each sentence with contextual information from prior sentences, and computing per-sentence relevance probabilities. Relevant sentences are extracted using bidirectional RNNs, feed-forward networks, Sparsemax activation, and CNNs to filter noise such as salutations and signatures. The extracted content is then matched against candidate agent responses to surface the most appropriate reply.

Working Papers

Active research that is not yet formally submitted — results and framing may evolve.

Working Paper2025
Working Paper

Context-Indexed Route Memory: Non-Destructive Storage and Compositional Execution of Incompatible Transition Systems

Paul Gradie

We study a controlled continual compositional route-memory benchmark where models must store and execute multiple incompatible transition systems over the same state and action space. The central question is whether context-indexed structural plasticity combined with recurrent execution supports non-destructive storage and compositional execution of such incompatible route systems. Across a suite of controlled experiments, we investigate the conditions under which models can store and retrieve distinct route memories without interference.

Theses

Doctoral and postgraduate research dissertations.

University of Melbourne2017
Doctoral Thesis

Defining the Master Regulator of Urethral Closure in Mouse

Paul Gradie

This doctoral dissertation investigates the molecular regulation of urethral closure in mouse, with the goal of identifying the master transcriptional regulator that drives the fusion of urethral folds during male development. Using genetic, genomic, and molecular approaches, the work characterises the gene regulatory networks governing a critical developmental process whose disruption causes hypospadias — one of the most common congenital anomalies in males.