Graham Wilcock

WikiTalk is basically a talking Wikipedia. You tell the robot what topics you are interested in, and the robot tells you about them using the information from Wikipedia. The robot gets the information directly from Wikpedia, so you get the latest up-to-date information even for rapidly changing topics. Wikipedia information is more trustworthy than some other sources, because it is contimuously validated by a worldwide community of volunteer editors.

WikiTalk anticipates what you will probably want to hear about next and extracts links to related topics in Wikipedia. When you say the name of a related topic, the robot switches to the new topic and starts talking about it. Predicting likely topic shifts like this also helps speech recognition.

WikiTalk is multimodal and multilingual. It integrates face-tracking, nodding and communicative gesturing with speech synthesis and speech recognition, following the CDM Constructive Dialogue Model. The system now works in English, Finnish and Japanese.

Brief history of WikiTalk

WikiTalk was developed without support from University of Helsinki. With no funding for robots, the first version of WikiTalk in 2011 ran on a Pyro (Python robotics) simulator. In 2012 WikiTalk on Nao robots was implemented at Supélec in Metz, France as an international PhD summer school project. A Finnish language localisation for Finnish-speaking robots was made by the Academy of Finland DigiSami project as a step towards SamiTalk, using Nao robots borrowed from SASKY in Sastamala.

Since 2016 WikiTalk has been developed by CDM Interact, a Finnish startup company. A Japanese language localisation for Japanese-speaking robots was made at Doshisha University, Japan. A demo of multilingual WikiTalk with language-switching on a Nao robot was presented at COLING 2016 (International Conference on Computational Linguistics) in Osaka, Japan. In 2017 WikiTalk won an award in Best Robot Design (Software Category) at ICSR 2017 (International Conference on Social Robotics) in Tsukuba, Japan.

In 2019 WikiTalk was used with ERICA the female android robot at Kyoto University, Japan where Graham Wilcock was Visiting Professor. A demo of ERICA and WikiTalk was presented at IJCAI 2019 (International Joint Conference on Artificial Intelligence) in Macao, China.

The DigiSami project was part of a wider joint research project funded by the Academy of Finland and the Hungarian Academy of Science, intended to increase the visibility and use of small Fenno-Ugric languages in the digital world by applying language technology to support the development of digital materials. The goal of the Finnish part was to support digital content generation for the Sami languages, focusing mainly on North Sami. The DigiSami project at University of Helsinki investigated how modern language technology and corpus-based linguistic research can contribute to handling the digitalisation challenges faced by small Fenno-Ugric language communities.

DigiSami Corpus

The project collected the DigiSami Corpus of Spoken North Sami language in several North Sami-speaking areas of both Finland and Norway. The spoken dialogues were transcribed and carefully annotated using modern corpus linguistics techniques. The speech materials were made available to colleagues at Aalto University working on speech technology, who began to develop a speech synthesizer and a speech recognizer for North Sami.

SamiTalk

The project worked towards SamiTalk, a Sami-speaking robot linked to Sami Wikipedia. The aim was to demonstrate to young Sami speakers that the Sami languages of their grandparents will also play an active part in the digital future. The WikiTalk project had shown that robots can talk about many topics using English Wikipedia. As a step towards SamiTalk, the DigiSami project worked on localisation of spoken dialogue systems and made a Finnish localisation of WikiTalk using Finnish Wikipedia. A SamiTalk demonstration prototype shows a robot talking about the situation of Sami languages in Finland using information from Sami Wikipedia. The robot speaks North Sami, but can only do speech recognition for Finnish as the North Sami speech recognizer was not available.

International Workshop

The project organised the northernmost International Workshop on Spoken Dialogue Systems in Saariselkä, Finland in 2016. The workshop attracted over 50 researchers from Europe, Japan and USA. Revised versions of the workshop papers were collected in Dialogues with Social Robots, edited by Kristiina Jokinen and Graham Wilcock and published by Springer in 2017.

Best Paper Award

The project also worked on multimodal analysis of the videos and annotations in the DigiSami Corpus using machine learning techniques, in collaboration with University of Eastern Finland. This work was recognised by a Best Paper Award at IWSDS 2018 for Enabling Spoken Dialogue Systems for Low-resourced Languages: End-to-end Dialect Recognition for North Sami by Trung Ngo Trong, Kristiina Jokinen and Ville Hautamäki.