Probabilistic Action Cores
{{Infobox
|bodystyle = width:30em|titlestyle = |title = PRAC|imagestyle = |captionstyle = |image = |caption = |headerstyle = |labelstyle = width:33%|datastyle = |header1 = |label1 = Version|data1 = 1.1.2|header2 = |label2 = Framework|data2 = Python|header3 = |label3 = Type|data3 = Natural-language instruction interpreter|header4 = |label4 = License|data4 = BSD|header5 = |label5 = Lead Developer|data5 = Daniel Nyga|header6 = |label6 = Institute|data6 = Institute for Artificial Intelligence, University of Bremen|header7 = |label7 = Website|data7 = http://www.actioncores.org|}}
PRAC (Probabilistic Action Cores) is an interpreter for natural-language instructions for robotic applications developed at the Institute for Artificial Intelligence at the University of Bremen, Germany, and is supported in parts by the European Commission and the German Research Foundation (DFG).{{cite journal |first1 = Daniel |last1 = Nyga |title = Interpretation of Natural-language Robot Instructions: Probabilistic Knowledge Representation, Learning, and Reasoning |journal = PhD Thesis | year = 2017 | url = http://nbn-resolving.de/urn:nbn:de:gbv:46-00105882-13 |format = PDF}}
Goals
The ultimate goal of the PRAC system is to make knowledge about everyday activities from websites like wikiHow available for service robots, such that they can autonomously acquire new high-level skills by browsing the Web.{{cite book |doi = 10.1109/IROS.2012.6385923 |citeseerx = 10.1.1.708.7035 |isbn = 978-1-4673-1736-8 |chapter = Everything robots always wanted to know about housework (But were afraid to ask) |title = 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems |pages = 243–250 |year = 2012 |last1 = Nyga |first1 = Daniel |last2 = Beetz |first2 = Michael |s2cid = 302048 }} PRAC addresses the problem that natural language is inherently vague and unspecific. To this end, PRAC maintains probabilistic first-order knowledge bases over semantic networks represented in Markov logic networks. As opposed to other semantic learning initiatives like NELL or IBM's Watson, PRAC does not aim at answering questions in natural language, but to disambiguate and infer information pieces that are missing in natural-language instructions, such that they can be executed by a robot. "This problem formulation is substantially different to the problem of text understanding for question answering or machine translation. In those reasoning tasks, the vagueness and ambiguity of natural-language expressions can often be kept and translated into other languages. In contrast, robotic agents have to infer missing information pieces and disambiguate the meaning of the instruction in order to perform the instruction successfully."{{cite journal |first1 = Daniel |last1 = Nyga |first2 = Michael |last2 = Beetz |title = Cloud-based Probabilistic Knowledge Services for Instruction Interpretation |journal = International Symposium of Robotics Research (ISRR) | year = 2015 | url = http://www.acat-project.eu/modules/BibtexModule/uploads/PDF/nygabeetz2015a.pdf }} In addition to probabilistic relational models, PRAC uses the principles of analogical reasoning and instance-based learning to infer completions of roles in semantic networks.{{cite journal |first1 = Daniel |last1 = Nyga |first2 = Mareike |last2 = Picklum | first3 = Sebastian | last3 = Koralewski | first4 = Michael |last4 = Beetz |title = Instruction Completion through Instance-based Learning and Semantic Analogical Reasoning |journal = International Conference on Robotics and Automation (ICRA) | year = 2017}}
PRAC has been successfully applied to teach robots to conduct chemical experiments{{cite book |doi = 10.1109/IROS.2015.7354110 |isbn = 978-1-4799-9994-1 |chapter = Towards robots conducting chemical experiments |title = 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) |pages = 5202–5208 |year = 2015 |last1 = Lisca |first1 = Gheorghe |last2 = Nyga |first2 = Daniel |last3 = Balint-Benczedi |first3 = Ferenc |last4 = Langer |first4 = Hagen |last5 = Beetz |first5 = Michael |s2cid = 7613082 }} and to make pancakes and pizza from wikiHow articles.{{cite news |author= Will Knight |title=Robots Learn to Make Pancakes from WikiHow Articles | url=https://www.technologyreview.com/s/540781/robots-learn-to-make-pancakes-from-wikihow-articles/ |quote=A robot called PR2 in Germany is learning to prepare pancakes and pizzas by carefully reading through WikiHow’s written directions. |work=MIT Tech Review |date= August 24, 2015 |accessdate=2017-03-14 }}
References
{{reflist}}
External links
- [http://www.actioncores.org Project homepage]
- [http://ai.uni-bremen.de Institute for Artificial Intelligence]
Category:Natural language processing software