Naomi Feldman
Department of Linguistics
1401 Marie Mount Hall
College Park, MD 20742
(301) 405-5800
nhf (at) umd (dot) edu

I'm an assistant professor in the Department of Linguistics at the University of Maryland with a focus in computational psycholinguistics and phonology. My research uses methods from statistics and machine learning to formalize questions about how people learn and represent the structure of their language. I use computational models to identify strategies that would allow people to process language effectively, and then I use behavioral experiments to determine which of these strategies people actually use. I have primarily applied these methods to studying speech sound category representations, investigating how people learn sound categories robustly from limited data and how those categories affect their subsequent perception of sounds.

projects | publications | links | fun

Recent and ongoing projects

Perceptual Warping
Data on the perceptual magnet effect (Kuhl, 1991) indicate that perceptual space is shrunk in the neighborhood of vowel prototypes and expanded near category boundaries. We are modeling this pattern using a statistical model in which listeners need to recover the phonetic detail of a speaker's target production through a noisy speech signal. Because experienced listeners know that speakers are more likely to produce sounds near the centers of phonetic categories, they should bias their perception toward category centers for optimal accuracy. We have shown a close correspondence between model predictions and data from Iverson & Kuhl (1995) and have collected data showing that as predicted, listeners use category information to different degrees depending on the amount of noise in the speech signal.

We have also applied this model to stop consonants and fricatives, testing whether the strong categorical effects found in consonants can be described as simple parametric variants of the weaker effects in vowels. Our hypothesis is that consonants have much less meaningful within-category variability (e.g., such as might arise through coarticulation), relative to the noise through which they are perceived, than do vowels. This should lead listeners to attend less to phonetic detail and to rely more heavily on knowledge of phonetic categories when perceiving consonants.

Feldman and Griffiths (2007) - conference paper introducing the Bayesian model
Feldman, Griffiths, and Morgan (2009) - paper with a full description of the Bayesian model and the noise experiment
Shi, Griffiths, Feldman, and Sanborn (2010) - paper describing an exemplar-based implementation of the model
Kronrod, Coppess, and Feldman (in press) - conference paper extending the model to consonant perception
Lexical Influences on Phonetic Category Acquisition
Infants begin to segment words from fluent speech at about the same time they acquire native language phonetic categories (between 6 and 12 months). A learner that simultaneously learns about both levels of structure might be able to use information about which sounds occur together in words when deciding whether sounds belong to the same or different phonetic categories. We are exploring how this type of information can provide useful constraints to guide phonetic category acquisition. Our simulations suggest that information from words can help disambiguate overlapping categories in cases where distributional information alone would not be sufficient. We have shown experimentally that adults use information from words to constrain their interpretation of phonetic variability, and a parallel experiment with infants is currently underway. This work is funded by NSF grant BCS-0924821.

Feldman, Griffiths, and Morgan (2009) - conference paper describing the model
Feldman, Myers, White, Griffiths, and Morgan (2011) - conference paper describing the experiment with adults

Publications

Gagliardi, A., Bennett, E., Lidz, J., & Feldman, N. H. (in press). "Children's inferences in generalizing novel nouns and adjectives." Proceedings of the 34th Annual Conference of the Cognitive Science Society.

Gagliardi, A., Feldman, N. H., & Lidz, J. (in press). "When suboptimal behavior is optimal and why: Modeling the acquisition of noun classes in Tsez." Proceedings of the 34th Annual Conference of the Cognitive Science Society.

Kronrod, Y., Coppess, E., & Feldman, N. H. (in press). "A unified model of categorical effects in consonant and vowel perception." Proceedings of the 34th Annual Conference of the Cognitive Science Society.

Feldman, N., Myers, E., White, K., Griffiths, T., & Morgan, J. (2011). "Learners use word-level statistics in phonetic category acquisition." Proceedings of the 35th Boston University Conference on Language Development.

Shi, L., Griffiths, T. L., Feldman, N. H., & Sanborn, A. N. (2010). "Exemplar models as a mechanism for performing Bayesian inference." Psychonomic Bulletin and Review, 17(4), 443-464. [Best Article of the Year Award]

Feldman, N. H., Griffiths, T. L., & Morgan, J. L. (2009). "The influence of categories on perception: Explaining the perceptual magnet effect as optimal statistical inference." Psychological Review, 116(4), 752-782.

Feldman, N. H., Griffiths, T. L., & Morgan, J. L. (2009). "Learning phonetic categories by learning a lexicon." Proceedings of the 31st Annual Conference of the Cognitive Science Society.

Soderstrom, M., Conwell, E., Feldman, N., & Morgan, J. (2009). "The learner as statistician: Three principles of computational success in language acquisition." Developmental Science, 12(3), 409-411.

Shi, L., Feldman, N. H., & Griffiths, T. L. (2008). "Performing Bayesian inference with exemplar models." Proceedings of the 30th Annual Conference of the Cognitive Science Society.

Feldman, N. H. & Griffiths, T. L. (2007). "A rational account of the perceptual magnet effect." Proceedings of the 29th Annual Conference of the Cognitive Science Society.

Related links

Language Science (IGERT program)
Machine Learning Reading Group
Infant Studies at Maryland

Metcalf Infant Research Lab (Brown)
Computational Cognitive Science Lab (Berkeley)

Non-academic

I did several a cappella arrangements, both Jewish and secular, while a member of Rhythm and Jews and Honorable Menschen. You are welcome to use them in concerts and recordings. Please give me credit if there is a printed program or insert (and I'd love to hear any recordings you make).
Below is the drumkit I use for vocal percussion. All percussion should be devoiced, especially for females. (Actually, most of the sounds are ejectives; take a phonetics course to find out what that means!)

dp bass drum
pf snare drum
k   rim shot on a snare  
tst high hat
ts-k repeated high hat
ksh cymbals
  frr (rolled 'r')   drum roll


I'm in a web comic! I play the role of "some girl" at Brown University.