Computational Cognitive Science

A list of potential topics for PhD students in the area of Computational Cognitive Science.

Neural Network Models of Human Language and Visual Processing

Supervisor: Frank Keller

Recent neural models have used attention mechanisms as a way of focusing the processing of a neural networks on certain parts of the input. This has proved successful for diverse applications such as image description, question answering, or machine translation. Attention is also a natural way of understanding human cognitive processing: during language processing, humans attend words in a certain order; during visual processing, they view image regions in a certain sequence. Crucially, human attention can be captured precisely using an eye-tracker, a device that measures which parts of the input the eye fixates, and for how long. Projects within this area will leverage neural attention mechanisms to model aspects of human attention. Examples include reading: when reading text, humans systematically skip words, spend more time on difficult words, and sometimes re-read passages. Another example is visual search: when looking for a target, human make a sequence of fixations which depend a diverse range of factors, such as visual salience, scene type, and object context. Neural attention models that capture such behaviors need to combine different types of knowledge, while also offering a cognitively plausible story how such knowledge is acquired, often based on only small amounts of training data.

Topics in morphology (NLP or cognitive modelling)

Supevisor:  Sharon Goldwater

Many NLP systems developed for English ignore the morphological structure of words and (mostly) get away with it. Yet morphology is far more important in many other languages. Handling morphology appropriately can reduce sparse data problems in NLP, and understanding human knowledge of morphology is a long-standing scientific question in cognitive science. New methods in both probabilistic modeling and neural networks have the potential to improve word representations for downstream NLP tasks and perhaps to shed light on human morphological acquisition and processing. Projects in this area could involve working to combine distributional syntactic/semantic information with morphological information to improve word representations for low-resource languages or sparse datasets, evaluating new or existing models of morphology against human behavioral benchmarks, or related topics.

Topics in unsupervised speech processing and/or modelling infant speech perception

Supervisor: Sharon Goldwater

Work in unsupervised (or 'zero-resource') speech processing (see Aren Jansen et al., ICASSP 2013) has begun to investigate methods for extracting repeated units (phones, words) from raw acoustic data as a possible way to index and summarize speech files without the need for transcription.  This could be especially useful in languages where there is little data to develop supervised speech recognition systems.  In addition, it raises the possibility of whether similar methods could be used to model the way that human infants begin to identify words in the speech stream of their native language.  Unsupervised speech processing is a growing area of research with many interesting open questions, so a number of projects are possible.  Projects could focus mainly on ASR technology or mainly on modeling language acquisition; specific research questions will depend on this choice.  here are just two possibilities: (1) unsupervised learners are more sensitive to input representation than are supervised learners, and preliminary work suggests that MFCCs are not necessarily the best option.  Investigate how to learn better input representations (e.g., using neural networks) that are robust to speaker differences but encode linguistically meaningful differences.  (2) existing work in both speech processing and cognitive modeling suggests that trying to learn either words or phones alone may be too difficult and in fact we need to develop *joint learners* that simultaneously learn at both levels.  Investigate models that can be used to do this and evaluate how joint learning can improve performance.

Communicative efficiency approaches to language processing/typology

Supervisor: Frank Mollica

In recent years, communicative efficiency has been formalized in terms of information theory and case studies (e.g., word order patterns, color-naming systems) have been used to demonstrate that linguistic forms and meanings support efficient communication. Despite having communication as a universal objective, the languages of the world are still highly diverse in both their communicative objectives and the strategies they use to achieve efficiency. Projects in this area would involve using information theoretic models and conducting experiments to: identify and characterize communicative functions and grammar strategies; predict and explain the prevalence of communicative functions and grammar strategies across cultures and groups; and investigate the developmental and evolutionary dynamics of grammar.

Modelling word learning, social learning and/or conceptual development

Supervisor: Frank Mollica

A fundamental goal of cognitive science is to understand how we construct and update rich mental models of the world both from experience with the world and from language. Development and evolution provide important constraints on our conceptual representations and cognitive technologies (e.g., language, math). As such, formal models of development proffer an explanatory framework for the patterns of behavior over the course of learning and at the “end state” of learning. Using developmental models, theoretical questions about the information that is stored by our concepts can be redefined as empirical questions of learning mechanisms, the environmental availability of data, the utility of data to the learner and inductive biases. Projects in this area could involve: building and evaluating interpretable models of human conceptual development; evaluating developmental and evolutionary constraints on conceptual representations; comparing different learning mechanisms (inductive vs pedagogical) against empirical data; building and evaluating models to account for the diversity of conceptual systems/cognitive technologies across individuals, groups and cultures.

Semantic / pragmatic processing strategies

Supervisor: Frank Mollica

Language allows us to achieve goals in the world and to build mental representations of entirely new worlds, complete with intricate economic, societal, cultural, moral and linguistic systems. Importantly, readers construct these rich mental models of a world through language and the representations they have built from using language to interact with the world. How do we use language to interface with the world and how do we learn and reason about worlds from linguistic input? There are often multiple strategies that we can adopt to make these updates, how do we choose which strategy to use in a given context? Projects in this area could include: investigating how we construct these representations and how they are updated during language processing; modelling how language interfaces with conceptual knowledge to accomplish goals (e.g., visual search); investigating how we build contextual representations for language processing and how context influences semantic/pragmatic processing. Projects will likely involve empirical data collection, Bayesian data analysis techniques, Bayesian cognitive modelling and program induction / reinforcement learning methods.