Laboratory of Advanced
speech and language Technology



About us


The Laboratory of Advanced speech and language Technology (the successor of the Laboratory of Advanced Speech and Language Technology of the AMU Foundation) is a research unit consisting of highly experienced specialists in science domains such as natural language processing and arftificial intelligence. The combination of extensive linguistic knowledge and implementing state-of-the-art technology solutions lays the basis for our research work. Our research work includes creating, maintaining and analysing speech corpora, professional speech and text annotation and developing advanced NLP modules. We provide AI solutions by conducting significant interdisciplinary projects gathering scientific experts. Our effective cooperation ensures the functionality and optimality of the results of our work. Our research work includes creating speech and text corpora related to extraction, access, interpretation and understanding of information (e.g. machine learning), speech and speaker recognition, speech synthesis, human-machine communication, linguistics, artificial intelligence and robotics. We are experienced in processing speech and language into four main areas of application of speech resources: general linguistic research activities, audiology and phoniatrics, speech technology as well as education. Some of our noticable achievements are: An application for automatic Polish language transcription (Implementation of Grapheme–to–Phoneme Rules and Extended SAMPA Alphabet in Polish Text–to–Speech Synthesis), Polish language synthesis (Polish unit selection speech synthesis with BOSS: extensions and speech corpora), Speech learning system AZAR (Applying speech and language technology to foreign language education), Audiometric tests for hearing rehabilitation (Linguistic and Phonetic Fundaments of Speech Tests), Polish speech corpus (Corpus based spoken language studies).


Our Mission


Research

We are constantly gathering vocal samples and creating speech corpora for research and development puproses.

Technology

With the help of scientists particularly skilled in artificial intelligence we provide optimal NLP solutions deployable in many purposes

Language

Our work includes various languages from many regions across the world


Projects


Current Projects


AI Searcher Automatic Analysis of Phonetic Convergence in Speech Technology Systems
AISearcher is an innovative system supporting the operation of analytical and check-in systems of the Border Guard. The system easily joins existing Border Guard systems. The AISercher controller will be an expert module based on artificial intelligence. AISercher will contain a set of necessary tools for analysing Internet resources and useful applications in the detection of offences registered by the Border Guard, typing suspects and criminal groups, and tracking the relationship between them. An essential advantage of AISercher is the module of automatic translation and statistical analysis of the meanings of individual words. Using linguistic methods, including morphological analysis, syntax and semantics of the analysed text, will be possible to work effectively on the text written in the following languages: Polish, Ukrainian, Russian and Belarusian. Heuristic algorithms, neural networks and large data sets (Big Data) will be used to search the Internet. This project focuses on the analysis and objective evaluation of phonetic convergence in human-to-human and human-to-machine conversations. The notion of phonetic convergence is related to the Communication Accommodation Theory (CAT) which regards interpersonal conversation as a dynamic adaptive exchange. The main objective of this project is to provide a quantitative description of accommodation phenomena exhibited in specific properties of speech (acoustic-prosodic, temporal and spectral) in human and human-computer dialogues in view of implementation in speech technology. In particular, it will be important to answer the question whether simultaneous use of rough sets method and neural networks provides a good theoretical basis for studying imprecision, vagueness and contradictions in the linguistic and phonetic analysis of data.

Past Projects


A high quality corpus based German–Polish speech synthesis Intelligent Language Tutoring System (ILTS) with multimodal feedback functions
Modelling of Prosody for Advanced Spoken Language Processing. Prosody modeling through inter–language analysis and its use for advanced CALL system

International and National
Cooperation


logo
usf
politechnika poz
logo
logo
logo


Our Expertise


Avatar
Language Analysis
-
Lexicons Building and Labelling / Language Resuorces Gathering with the Use of Text Processing
Avatar
Interdisciplinarity
-
Linguistic Analysis for Medical and Sociological Research
Avatar
AI Development
-
Creating Machine Learning Models and Algorithms
Avatar
Speech Analysis
-
Acoustic and Phonetic Analysis and Speech Annotation / Speech Processing - Recognition and Synthesis, Speaker Recognition


Contact


ul. Uniwersytetu Poznańskiego 10
61-614 Poznań, Poland

wcat.langlab@gmail.com

Tel. (+4861)829 1885