In this paper, we discuss a computational model that is able to detect and build word-like representations on the basis of sensory input. The model is designed and tested with a further aim to investigate how infants may learn to communicate by means of spoken language. The computational model makes use of a memory, a perception module, and the concept of ’learning drive’. Learning takes place within a communicative loop between a ’caregiver’ and the ’learner’. Experiments carried out on three European languages with different genetic background (Finnish, Swedish, and Dutch) show that a robust word representation can be learned in using less than 100 acoustic tokens (examples) of that word. The model is inspired by the memory structure that is assumed functional for human cognitive processing.

Hidden Markov Models have been essential in obtaining today’s successes in speech recognition. However, some limitations of HMMs become clear: for example it is difficult to successfully exploit features that are measured at different time scales than the centisecond scale at which the spectral features are measured. Little success has been achieved in integrating utterance level information such as prosody, segmental information and finer detail such as voice onset times. In this paper, we apply latent semantic analysis (LSA) techniques known from the text processing field to histograms of acoustic event co-occurrence (HAC) to propose a novel speech recognition framework. We show that the HACmethod can deal with correlated information and exploit knowledge sources that are asynchronous. Index Terms: speech recognition, information discovery, information integration, latent semantic analysis, cooccurrence statistics. 1.

The multigram algorithm is a statistical technique that can be used for extracting recurring patterns from a sequential input. When provided with a symbol sequence representing a speech signal, it is able to extract word-like patterns from it, despite the large amount of subsequences that can represent a single word. For this, it uses statistical information derived from the entire input. However, due to the abstraction of speech to symbols, much of the information originally present in the signal is no longer available to the algorithm. In this paper we propose a way of using a richer abstraction of the signal in the form of a lattice. Furthermore, a way of grounding recurring patterns to concepts in other modalities will be presented. Finally, the information learned by the algorithm using both kinds of input is tested in a recognition experiment. This will show that the use of lattices leads to a significant improvement in terms of recognition rate.

Abstract not found

In this paper, we discuss a computational model of language acquisition which focuses on the detection of words and that is able to detect and build word-like representations on the basis of multimodal input data. Experiments carried out on three European languages (Finnish, Swedish, and Dutch) show that internal word representations can be learned without a predefined lexicon. The computational model is inspired by a memory structure that is assumed to be functional for human cognitive processing. The model does not use any prior segmentation, nor does it use the concept of segmentation later in the processing. This calls into question the importance that is conventionally attributed to the segmentation of the speech signal in terms of symbolic units for the purpose of detecting structure in speech.

Abstract not found

In this paper we investigate a computational model of word learning that is cognitively plausible. The model is partly trained on incorrect form-referent pairings, modelling the input to a word-learning child that may contain such mismatches due to inattention to a joint communicative scene. We introduce a procedure of active learning, based on attested cognitive processes. We then show how this procedure can help overcome the unreliability of the input by detecting and correcting the mismatches by reliance on previously built up experience. Index Terms: language acquisition, word learning, computational modelling