Home of Derek Monner

[CV] [Resume]

Research Interests

Since Summer 2006 I've been working with Jim Reggia and the Biologically-Inspired Computing Group, where my research concerns neural networks and approaches to biologically-plausible models of natural language acquisition. I spent two recent years as an NSF IGERT Fellow to working in UMCP's Biological and Computational Foundations of Language Diversity program, and as such have been working closely with the Linguistics, Neuroscience and Cognitive Science, and other departments. Most recently I've rejoined Jim Reggia's lab as a postdoc working on more mainstream machine learning problems such as neural network approaches to classification in complex networked datasets.

As of early 2011, I have a major hand in five ongoing projects:

  • Researching novel neural network methods for approaching collective and relational classification problems;
  • Attempting to ground symbolic cognition in neural networks, both theoretically and experimentally;
  • My post-dissertation research on neural networks that learn to recognize, produce, and ground subsets of natural languages;
  • Using neural networks to model critical periods in the acquisition of gender agreement and assignment in second languages, where I work with an academically diverse team including Giovanna Morini, Karen Vatz, So-One Hwang, and Robert DeKeyser; and,
  • Using Twitter to predict real-world events, working with Bill Rand as part of the Social Media Micro Modeling group.

I have previously worked:

Below is a partial list of publications; for the full deal, see my CV.

Journal Publications

D. Monner and J. A. Reggia. Recurrent neural collective classification (2013). IEEE Transactions on Neural Networks and Learning Systems, 24(12), 1932--1943.[pdf]
D. Monner, K. Vatz, G. Morini, S. Hwang, and R. DeKeyser (2013). A neural network model of the effects of entrenchment and memory development on grammatical gender learning. Bilingualism: Language and Cognition 16(2), 246--265.[pdf]
D. Monner and J. A. Reggia (2012). Emergent latent symbol systems in recurrent neural networks. Connection Science 24(4), 193--225.[pdf]
D. Monner and J. A. Reggia (2012). Neural architectures for learning to answer questions. Biologically Inspired Cognitive Architectures 2, 37--53.[pdf]
D. Monner and J. A. Reggia (2012). A generalized LSTM-like training algorithm for second-order recurrent neural networks. Neural Networks 25, pp70-83.[pdf]

Conference Publications

J. A. Reggia, D. Monner, and J. Sylvester (2013). The computational explanatory gap. In Proceedings of IACAP.[pdf]
A. Grushin, D. Monner, J. A. Reggia, and A. Mishra. Robust human action recognition via long short-term memory. In Proceedings of IJCNN.[pdf]
D. Monner and J.A. Reggia (2011). Towards a biologically inspired question-answering neural architecture. In Proceedings of BICA. Amsterdam: IOS Press.[pdf]
D. Monner and J.A. Reggia (2011). Systematically grounding language through vision in a deep, recurrent neural network. In Proceedings of AGI. Springer.[pdf]
D. Monner and J.A. Reggia (2009). An unsupervised learning method for representing simple sentences. In Proceedings of IJCNN.[pdf]
A. Bender, R. Sherwood, D. Monner, N. Goergen, N. Spring, and B. Bhattacharjee (2009). Fighting spam with the NeighborhoodWatch DHT. In Proceedings of INFOCOM.[pdf]

Technical Reports

D. Monner (2011). A Neurocomputational Model of Grounded Language Comprehension and Production at the Sentence Level. Ph.D. Dissertation, University of Maryland, College Park. Winner of the 2012 Larry S. Davis Doctoral Dissertation Award.[pdf]
D. Monner and J. A. Reggia (2007). An external tabletop environment for an interactive brain model. CS-TR-4883/UMIACS-TR-2007-41, Dept. of Computer Science, University of Maryland, College Park.[pdf]