Efficient decoding with generative score-spaces using the expectation semiring

State-of-the-art speech recognisers are usually based on hidden Markov models (HMMs). They model a hidden symbol sequence with a Markov process, with the observations independent given that sequence. These assumptions yield efficient algorithms, but limit the power of the model. An alternative model that allows a wide range of features, including word- and phone-level features, is a log-linear model. To handle, for example, word-level variable-length features, the original feature vectors must be segmented into words. Thus, decoding must find the optimal combination of segmentation of the utterance into words and word sequence. Features must therefore be extracted for each possible segment of audio. For many types of features, this becomes slow. In this paper, long-span features are derived from the likelihoods of word HMMs. Derivatives of the log-likelihoods, which break the Markov assumption, are appended. Previously, decoding with this model took cubic time in the length of the sequence, and longer for higher-order derivatives. This paper shows how to decode in quadratic time. © 2013 IEEE.

Commercializing a disruptive technology based upon university ip through open innovation

This paper discusses the use of a university spin-out firm to bring a potentially disruptive technology to market. The focus for discussion is how a spin-out can build a technology ecosystem of providers of complementary resources to enable partner organizations to build competence in a novel and potentially disruptive technology. The paper uses the illustrative case of Cambridge Display Technology Ltd (CDT) to consider these issues from the perspective of the literature on open innovation (with particular emphasis on the role of partnerships between start-ups and established firms), the commercialization of university IP, and the commercialization of disruptive technologies. © World Scientific Publishing Company.

Negative Effects of Relative Proximity and Absolute Geography on Open Innovation Practices in High-tech SMEs in the UK

© 2014 IEEE. This exploratory study addresses a gap in management literature by addressing the role of location in the continuously expanding field of open innovation research. In this context, we analyze potential negative effects of absolute geography and relative proximity on open innovation practices in high-tech small and medium-sized enterprises (SMEs) in the United Kingdom. Drawing upon cluster theory and business ecosystem literature, the analysis from three SME case studies in the East of England suggests that presumed 'favorable' location variables, such as close relative proximity between partners and the presence of economic clusters, can have certain negative effects on open innovation practices.