No. 057 Towards Explanation Production Combining Natural Language Processing and Logical Reasoning

Randy Goebel, University of Alberta, rgoebel@ualberta.ca

Yuji Matsumoto, NAIST, matsu@is.naist.jp

Ken Satoh, NII and Sokendai, ksatoh@nii.ac.jp

 

Akiko Aizawa, National Institute of Informatics, Japan, aizawa@nii.ac.jp
Yu Asano, Central Research Laboratory, Hitachi, Ltd., Japan, yu.asano.ko@hitachi.com
Chitta Baral,? Arizona State UniversityUSA, chitta@asu.edu
Bart Verheji, University of Groningen, b.verheij@ai.rug.nl
Ido Dagan, Bar-Ilan University, Israel, dagan@cs.biu.ac.il
Kentaro Inui, Tohoku University, Japan, inui@ecei.tohoku.ac.jp
Robert Kowalski, Imperial College London, UK, r.kowalski@imperial.ac.uk
Sadao Kurohashi, Kyoto University, Japan, kuro@i.kyoto-u.ac.jp
Bernardo Magnini, Fundazione Bruno Kessler(FBK), Italy, magnini@fbk.eu
Yusuke Miyao, National Institute of Informatics, Japan, yusuke@nii.ac.jp
Guenter Neumann, DFKI, Germany, neumann@dfki.de
Minh Le Nguyen, JAIST, Japan, nguyenml@jaist.ac.jp
Erich Schweighofer, University of Vienna, Austria, erich.schweighofer@univie.ac.at
Yuzuru Tanaka, Hokkaido University, Japan, tanaka@meme.hokudai.ac.jp
Satoshi Tojo, JAIST, Japan, tojo@jaist.ac.jp
Bart Verheij, University of Groningen, Netherland, b.verheij@ai.rug.nl
Akihiro Yamamoto, Kyoto University, Japan, akihiro@i.kyoto-u.ac.jp

The recent application of high performance natural language processing (NLP) systems has shifted from general question answering to more targeted application domains (e.g., medical diagnosis, epidemiology, legal reasoning, intellectual property management). Examples include the WATSON project by IBM and the Todai Robot Project by NII. But these recent systems have focused more on engineering problem solutions, and less so on general architectures for building domain models that support more general representation and reasoning within the specific application domains. For example, the WATSON architect uses an associative pattern matching process which supports high performance, but is difficult to follow for humans, and thus prevents easy collaborative problem-solving with humans. Our workshop proposal if focused on the development and integration of modern natural language processing tools to support the development of systems with methods of hypothesis management and explanation. In this way, humans can collaborate with such systems to not only understand the use of representation to create answers, but an also support incremental supervised machine learning. The purpose of the proposed meeting is to gather a group of researchers in related areas, such as natural language processing, information extraction, and logical reasoning, in order to formulate approaches to combining these areas to achieve the goal of machine-human collaboration in high-performance natural language-based domain interactive reasoning systems.

A research agenda for such a meeting would include a variety of topics, including language-based information extraction (e.g., open information extraction), reasoning architectures based on abduction and hypothesis management, natural language entailment, and inductive learning.

Specific challenges include the capture and use of legal documents, in order to answer legal questions. In this context, the creation of legal judgements involves not just answers (e.g., guilty, not guilty), but articulation of elaborate explanations supported by interpretation of legal statues and regulations. To construct such logical explanations, we need to identify information relevant to the questions, ensure that information is transformed to or directly represented in formal representations that support the contraction of explanations. In this phase, we need a NLP analysis of question and information retrieval technique based on the analysis. Subsequent steps include the development of entailment testing methods, to decide whether and how (or how not) a question is entailed by the logical representation of the extracted domain information. The general development of NLP techniques for this kind of subsumption involves ontological manipulation, relation extraction, and logical reasoning.

Because this diversity of requirements, our workshop will need to include researchers from information retrieval, information extraction, question-answer, general natural language processing, and logical representation and reasoning. With this context, the goal is to develop a basis for reliable system which produces an explanation of system’s behaviour, and be incrementally constructed and improved to advance their level of performance in a variety of appellation domains.

Possible topics of the meeting include, but are not limited to:

• Natural Language Processing for Information Retrieval
• Information Retrieval for Complex Logical Formula
• Natural Language Processing for Textual Entailment
• Natural Language Processing for Subsumption Test
• Ontology Research for Subsumption Test
• Combining Logical Reasoning and Natural Language Processing
• Feedback of Logical Reasoning to Solve Disambiguation

General information

• Japan Travel Guide
• Guide to living in Japan (PDF)
  Ministry of Foreign Affairs guide with lots of helpful information
• JR East great deals
• Suica card

Getting to Shonan Village

The following instructions cover the common cases for a non-resident of Japan to move from Tokyo and its airports to Shonan Village. In general, first take trains to either JR’s Zushi station or Keikyu’s Shin-Zushi station, then take a bus or taxi. The Shonan Village Web site gives more information, but please do not hesitate to ask the organizers for help!

If you get lost

Railway staff (especially at ticketing offices) tend to speak some English and know about destinations. Bus drivers, maybe less likely.

Step 1 of 2: tracks

You can search for train schedules online. However, if using the “N’EX”, please focus on trains labeled “Narita Express” or “JR” and disregard the fares calculated.

From Tokyo’s Narita airport

Take JR trains to Zushi. JR is Japan’s “national” railway company. At the airport after exiting customs, go downstairs and find a JR office (not a Keisei office). Buy the N’EX
TOKYO Direct Ticket (One-way) (it saves you more than 2000 yen). Take “N’EX” (Narita Express) to Yokohama or Ofuna (better), then take the Shonan-Shinjuku or Yokosuka line to Zushi. The N’EX ticket will cover the JR train trip all the way to Zushi (about 2 hours) and, if you buy the round-trip N’EX ticket, back. Tell the JR representative that you’re going to Zushi and you know you have to change trains at Yokohama or Ofuna. The Narita Express train splits en route so board the car of your assigned seat.

From Tokyo’s Haneda airport

Warning! Trains in Japan do not run at night. Night buses are rare. If you come to Haneda airport at about 10pm (which is typical for many US flights), you will be stuck in Tokyo for the night and should plan accordingly (e.g., make a reservation for a hotel in the vicinity of Haneda or the Tokyo station — which is as far as you could possibly go.) It’s best to plan to arrive to Japan before 6-7pm.

Take Keikyu trains (not the monorail) to Shin-Zushi. Keikyu is one of Japan’s many private railway companies. At the airport, buy a PASMO or Suica card with about 2000 yen. (You might get a slightly better deal if you buy a PASMO card from a Keikyu ticket office and tell them you are going to Shin-Zushi.) Use the card to go to Shin-Zushi terminal (about 1 hour), possibly changing trains at Keikyu-Kamata and/or Kanazawa-Hakkei.

From central Tokyo

Take JR’s Yokosuka line (e.g., from Tokyo or Shinagawa station) or Shonan-Shinjuku line (e.g., from Ikebukuro, Shinjuku, or Shibuya) to Zushi (about 1 hour). You can also take Keikyu trains (e.g., from Shinagawa) to Shin-Zushi.

Step 2 of 2: roads

A Keikyu bus departs once or twice per hour from Zushi and Shin-Zushi stations (bay 1) to Shonan Village (the last stop). It takes 30 minutes and costs 350 yen. The bus time table is online. The departure times from Shin-Zushi are 2 minutes after the departure times from Zushi.

Pay for the bus with your Suica or PASMO card by touching it to the blue “IC” panel when you get on (through the center door) and again when you get off (through the front door). It is when you get off that you pay, so don’t ask the driver for a ticket when you get on. After you get off, your destination building is right across the road (north, in other words).

A taxi from Zushi or Shin-Zushi station to Shonan Village takes 20 minutes and costs 2500 or 3000 yen.

Travel tips

Arranging for a ride together

Please add a comment to this post to announce your arrival information and arrange to share transportation of taxi (I recommend to take a taxi at Zushi JR station since there are a pool of taxi there). The comments will be deleted after the seminar.

Earthquakes

If you have never experienced even a small earthquake, you may wish to check the page “What should I do when…?” just in case (just as you should know how to evacuate in case of a fire even if the chance is very small).

Meeting points in Tokyo

Haneda airport, international terminal: after exiting baggage claim and customs, turn right and you’ll see the tourist information office to your right. There are plenty of seats in front of the office.

Narita airport: please see http://www.narita-airport.jp/en/guide/service/list/svc_29.html

JR Tokyo station, underground within the ticket gates: There are two meeting points. The “Gin-no-suzu” (a large silver bell) is traditional and next to many shops and services. It is on the first underground floor. http://www.tokyoinfo.com/guide/station/ginnosuzu.html Another place on the same floor is a big square in front of a relief steam locomotive structure. http://www.tokyoinfo.com/guide/station/#meet02

JR Ueno station: the ticket gates at the exit to the park, which is a good place to let daylight attenuate your jetlag. http://www.joho.st/tokyo/ueno/koen.php

You can also access to FAQ page about travel information.

This content is password protected. To view it please enter your password below:

Password: