A Corpus-Based Approach to Speech Enhancement from Nonstationary Noise

Temporal dynamics and speaker characteristics are two important features of speech that distinguish speech from noise. In this paper, we propose a method to maximally extract these two features of speech for speech enhancement. We demonstrate that this can reduce the requirement for prior information about the noise, which can be difficult to estimate for fast-varying noise. Given noisy speech, the new approach estimates clean speech by recognizing long segments of the clean speech as whole units. In the recognition, clean speech sentences, taken from a speech corpus, are used as examples. Matching segments are identified between the noisy sentence and the corpus sentences. The estimate is formed by using the longest matching segments found in the corpus sentences. Longer speech segments as whole units contain more distinct dynamics and richer speaker characteristics, and can be identified more accurately from noise than shorter speech segments. Therefore, estimation based on the longest recognized segments increases the noise immunity and hence the estimation accuracy. The new approach consists of a statistical model to represent up to sentence-long temporal dynamics in the corpus speech, and an algorithm to identify the longest matching segments between the noisy sentence and the corpus sentences. The algorithm is made more robust to noise uncertainty by introducing missing-feature based noise compensation into the corpus sentences. Experiments have been conducted on the TIMIT database for speech enhancement from various types of nonstationary noise including song, music, and crosstalk speech. The new approach has shown improved performance over conventional enhancement algorithms in both objective and subjective evaluations.

Strudel: A corpus-based semantic model based on properties and types

Computational models of meaning trained on naturally occurring text successfully model human performance on tasks involving simple similarity measures, but they characterize meaning in terms of undifferentiated bags of words or topical dimensions. This has led some to question their psychological plausibility (Murphy, 2002; Schunn, 1999). We present here a fully automatic method for extracting a structured and comprehensive set of concept descriptions directly from an English part-of-speech-tagged corpus. Concepts are characterized by weighted properties, enriched with concept-property types that approximate classical relations such as hypernymy and function. Our model outperforms comparable algorithms in cognitive tasks pertaining not only to concept-internal structures (discovering properties of concepts, grouping properties by property type) but also to inter-concept relations (clustering into superordinates), suggesting the empirical validity of the property-based approach. Copyright © 2009 Cognitive Science Society, Inc. All rights reserved.

Speech Enhancement from Additive Noise and Channel Distortion - a Corpus-Based Approach

This paper presents a new approach to single-channel speech enhancement involving both noise and channel distortion (i.e., convolutional noise). The approach is based on finding longest matching segments (LMS) from a corpus of clean, wideband speech. The approach adds three novel developments to our previous LMS research. First, we address the problem of channel distortion as well as additive noise. Second, we present an improved method for modeling noise. Third, we present an iterative algorithm for improved speech estimates. In experiments using speech recognition as a test with the Aurora 4 database, the use of our enhancement approach as a preprocessor for feature extraction significantly improved the performance of a baseline recognition system. In another comparison against conventional enhancement algorithms, both the PESQ and the segmental SNR ratings of the LMS algorithm were superior to the other methods for noisy speech enhancement. Index Terms: corpus-based speech model, longest matching segment, speech enhancement, speech recognition

Statistically-constrained shallow text marking:Techniques, evaluation paradigm and results

We present three natural language marking strategies based on fast and reliable shallow parsing techniques, and on widely available lexical resources: lexical substitution, adjective conjunction swaps, and relativiser switching. We test these techniques on a random sample of the British National Corpus. Individual candidate marks are checked for goodness of structural and semantic fit, using both lexical resources, and the web as a corpus. A representative sample of marks is given to 25 human judges to evaluate for acceptability and preservation of meaning. This establishes a correlation between corpus based felicity measures and perceived quality, and makes qualified predictions. Grammatical acceptability correlates with our automatic measure strongly (Pearson's r = 0.795, p = 0.001), allowing us to account for about two thirds of variability in human judgements. A moderate but statistically insignificant (Pearson's r = 0.422, p = 0.356) correlation is found with judgements of meaning preservation, indicating that the contextual window of five content words used for our automatic measure may need to be extended. © 2007 SPIE-IS&T.

Generating a Word-Emotion Lexicon from #Emotional Tweets

Research in emotion analysis of text suggest that emotion lexicon based features are superior to corpus based n-gram features. However the static nature of the general purpose emotion lexicons make them less suited to social media analysis, where the need to adopt to changes in vocabulary usage and context is crucial. In this paper we propose a set of methods to extract a word-emotion lexicon automatically from an emotion labelled corpus of tweets. Our results confirm that the features derived from these lexicons outperform the standard Bag-of-words features when applied to an emotion classification task. Furthermore, a comparative analysis with both manually crafted lexicons and a state-of-the-art lexicon generated using Point-Wise Mutual Information, show that the lexicons generated from the proposed methods lead to significantly better classi- fication performance.

A voxel based morphometry study investigating brain structural changes in first episode psychosis.

Schizophrenia (SCZ) and bipolar disorder (BP) are associated with neuropathological brain changes, which are believed to disrupt connectivity between brain processes and may have common properties. Patients at first psychotic episode are unique, as one can assess brain alterations at illness inception, when many confounders are reduced or absent. SCZ (N=25) and BP (N=24) patients were recruited in a regional first episode psychosis MRI study. VBM methods were used to study gray matter (GM) and white matter (WM) differences between patient groups and case by case matched controls. For both groups, deficits identified are more discrete than those typically reported in later stages of illness. SCZ patients showed some evidence of GM loss in cortical areas but most notable were in limbic structures such as hippocampus, thalamus and striatum and cerebellum. Consistent with disturbed neural connectivity WM alterations were also observed in limbic structures, the corpus callosum and many subgyral and sublobar regions in the parietal, temporal and frontal lobes. BP patients displayed less evidence of volume changes overall, compared to normal healthy participants, but those changes observed were primarily in WM areas which overlapped with regions identified in SCZ, including thalamus and cerebellum and subgyral and sublobar sites. At first episode of psychosis there is evidence of a neuroanatomical overlap between SCZ and BP with respect to brain structural changes, consistent with disturbed neural connectivity. There are also important differences however in that SCZ displays more extensive structural alteration.

Volume reduction of the corpus callosum and its relationship with deficits in interhemispheric transfer of information in recent-onset psychosis

The present study aimed to investigate the presence of corpus callosum (CC) volume deficits in a population-based recent-onset psychosis (ROP) sample, and whether CC volume relates to interhemispheric communication deficits. For this purpose, we used voxel-based morphometry comparisons of magnetic resonance imaging data between ROP (n = 122) and healthy control (n = 94) subjects. Subgroups (38 ROP and 39 controls) were investigated for correlations between CC volumes and performance on the Crossed Finger Localization Test (CFLT). Significant CC volume reductions in ROP subjects versus controls emerged after excluding substance misuse and non-right-handedness. CC reductions retained significance in the schizophrenia subgroup but not in affective psychoses subjects. There were significant positive correlations between CC volumes and CFLT scores in ROP subjects, specifically in subtasks involving interhemispheric communication. From these results, we can conclude that CC volume reductions are present in association with ROP. The relationship between such deficits and CFLT performance suggests that interhemispheric communication impairments are directly linked to CC abnormalities in ROP. (C) 2010 Elsevier Ireland Ltd. All rights reserved.

The role of imitation in video-based interventions for children with autism.

Objective: The aim of this paper is to bridge the gap between the corpus of imitation research and video-based intervention (VBI) research, and consider the impact imitation skills may be having on VBI outcomes and highlight potential areas for improving efficacy.

Method: A review of the imitation literature was conducted focusing on imitation skill deficits in children with autism followed by a critical review of the video modelling literature focusing on pre-intervention assessment of imitation skills and the impact imitation deficits may have on VBI outcomes.

Results: Children with autism have specific imitation deficits, which may impact VBI outcomes. Imitation training or procedural modifications made to videos may accommodate for these deficits.

Conclusions: There are only six studies where VBI researchers have taken pre-intervention imitation assessments using an assortment of imitation measures. More research is required to develop a standardised multi-dimensional imitation assessment battery that can better inform VBI.

Interpretable semantic vectors from a joint model of brain-and text-based meaning

Vector space models (VSMs) represent word meanings as points in a high dimensional space. VSMs are typically created using a large text corpora, and so represent word semantics as observed in text. We present a new algorithm (JNNSE) that can incorporate a measure of semantics not previously used to create VSMs: brain activation data recorded while people read words. The resulting model takes advantage of the complementary strengths and weaknesses of corpus and brain activation data to give a more complete representation of semantics. Evaluations show that the model 1) matches a behavioral measure of semantics more closely, 2) can be used to predict corpus data for unseen words and 3) has predictive power that generalizes across brain imaging technologies and across subjects. We believe that the model is thus a more faithful representation of mental vocabularies.