Unsupervised learning is a challenging task due to the lack of labels. Multiple Object Tracking (MOT), which inevitably suffers from mutual object interference, occlusion, etc., is even more difficult without label supervision. In this paper, we explore the latent consistency of sample features across video frames and propose an Unsupervised Contrastive Similarity Learning method, named UCSL, including three contrast modules: self-contrast, cross-contrast, and ambiguity contrast. Specifically, i) self-contrast uses intra-frame direct and inter-frame indirect contrast to obtain discriminative representations by maximizing self-similarity. ii) Cross-contrast aligns cross- and continuous-frame matching results, mitigating the persistent negative effect caused by object occlusion. And iii) ambiguity contrast matches ambiguous objects with each other to further increase the certainty of subsequent object association through an implicit manner. On existing benchmarks, our method outperforms the existing unsupervised methods using only limited help from ReID head, and even provides higher accuracy than lots of fully supervised methods.

Scientific WorkFlows (SWFs) play significant roles in scientific research and engineering simulation, which are often data intensive and has complex data dependencies. The storage of massive intermediate datasets has great impacts on the performance and the quality of service of a SWF system, which has become a difficult and complex task. Through analyzing the cost transitive tournament shortest path (CTT-SP)-based algorithm proposed for the intermediate data storage in SWF systems, the disadvantage of the CTT-SP algorithm over its sensitivity to the main branches can be found. We proposed an improved CTT-SP algorithm based on the critical path method (CPM), aiming at reducing the sensitivity of the main branch and improving the performance of the algorithm. By adding the generation information of the datasets on the arrows, the data dependency graph of a SWF can be converted to an activity on arrow (AOA) net. Then, the critical path of the AOA net is used to be the main branch of the CTT-SP algorithm, which can reduce the impacts of the main branch and keep the quality of service of the SWF systems. Experiments are designed to test the impacts of the main branches and the performance of the improved CTT-SP algorithms. The results show the significant impacts of the main branches on the performance of the CTT-SP algorithm and the effectiveness of the CPM on improving the performance of the CTT-SP algorithm. Comparison results demonstrate the positiveness and effectiveness of the improved CTT-SP algorithm.

This paper describes recent advances at LIMSI in Mandarin Chinese speech-to-text transcription. A number of novel approaches were introduced in the different system components. The acoustic models are trained on over 1600 hours of audio data from a range of sources, and include pitch and MLP features. N-gram and neural network language models are trained on very large corpora, over 3 billion words of texts; and LM adaptation was explored at different adaptation levels: per show, per snippet, or per speaker cluster. Character-based consensus decoding was found to outperform word-based consensus decoding for Mandarin. The improved system reduces the relative character error rate (CER) by about 10% on previous GALE development and evaluation data sets, obtaining a CER of 9.2% on the P4 broadcast news and broadcast conversation evaluation data.

This paper presents a method for generating indexable and browsable keyword metadata from ASR transcripts by leveraging the Web. Search engine queries are built from an ASR transcript and used to retrieve similar text from the Web. The keyword meta information embedded in those pages for search engines is then ranked using a mutual information criteria to derive a keyword set. The proposed method is training-free, allows phrase keyword generation, and can generate words that were not spoken in the ASR transcript, alleviating the impact of ASR out-of-vocabulary. Subjective evaluations on technical presentations demonstrate a clear preference for this approach. Additionally an objective measure of keyword generation performance is proposed and shown to be a useful guide for tuning compared to more onerous subjective evaluations.

This paper investigates unsupervised vocabulary and language model self-adaptation (VLA) from just one speech file using the web as a knowledge source and without prior knowledge of topic or domain beyond optional file metadata. Single-file self adaptation is regularly used for acoustic adaptation, but to date, is rarely used for VLA. The method investigated here uses a first-pass transcript or file metadata to generate web search queries for retrieving texts for adaptation. Various strategies for building queries, retrieving web texts and maximizing out-of-vocabulary (OOV) recovery while constraining vocabulary growth are examined. Significant improvements are demonstrated for transcribing and searching recorded lectures and telephone calls. The proposed method is orthogonal with acoustic adaptation and system combination and integrates well in multi-pass recognition architectures.

We examine the task of spoken term detection in Chinese spontaneous speech with a lattice-based approach. We first compare lattices generated with different units: word, character, tonal and toneless syllables, and also lattices converted from one unit to another unit. Then we combine lattices from multiple systems into a single lattice. By fully exploiting the redundant information in the combined lattice with a time-based node/arc merging, we achieve the result of a compact lattice index with the accuracy improved to 79.2% from 73.9% using the best subsystem.

We examine the task of spoken term detection in Chinese spontaneous speech with a lattice-based approach. We compare lattices generated with different units: word, character, tonal syllable and toneless syllable, and also look into methods of converting lattices from one unit to another one. We find the best system is with toneless-syllable lattices converted from word lattices. Further improvement is achieved by lattice post-processing and system combination. Our best system has an accuracy of 80.2% on a keyword spotting task.