Answer-Robotic-SemanticWeb-IoT-Workshop-IROS2019

From Knoesis wiki
Revision as of 16:59, 30 October 2019 by Amelie (Talk | contribs) (Program)

Jump to: navigation, search

3rd International Workshop on the Applications of Knowledge Representation and Semantic Technologies in Robotics (AnSWeR19).

Co-located with IROS 2019 (IEEE/RSJ International Conference on Intelligent Robots and Systems), A-rank conference.

Co-located with IROS

Flyer

International Workshop on the Applications of Knowledge Representation and Semantic Technologies in Robotics (AnSWeR19), co-located with IROS 2019 (IEEE/RSJ International Conference on Intelligent Robots and Systems)

Description of the AnSWeR19 Workshop

Autonomous mobile agents and robotics, in general, are experiencing a growing interest due to a number of factors, e.g. the advancements in Artificial Intelligence, Natural Language Processing, and Computer Vision; the amount of new efficient techniques for basic robotic tasks (perception, manipulation, navigation etc.); and the increasing number of cost-accessible robotic platforms in the market. As a consequence, robots will be required to achieve more and more complex tasks, hence exposing the ability to deal with different sources of knowledge about the world in order to improve their behaviors.

While the problem of enabling robots to use available sources of heterogeneous knowledge has attracted attention relatively recently in the robotics community (e.g. the RoboEarth or RoboBrain projects), the Knowledge Representation and Semantic Web communities have been studying techniques to model, integrate and exploit heterogeneous sources knowledge for a long time. We argue that such techniques should also have a role to play in the context of robotic applications, as they could be beneficial for robots to achieve their tasks. The goal It is therefore important to understand how these communities are interfacing, and how they can benefit from each other.

The goal of the AnSWeR workshop is to address these issues by studying the involvement and applications of Knowledge Representation formalisms and semantic technologies in robotic applications, including the emerging “Internet of Cloud Robotic Things” topic, by giving the researchers and practitioners the opportunity to compare and debate on common problems. The AnSWeR workshop will, therefore, allow addressing and debating on problems that have been tackled so far by two communities that worked on overlapping topics.

The first two editions of AnSWeR were held in 2017 and 2018 in Semantic Web conferences to investigate the interests in establishing synergies between both sides. An exciting dialogue between roboticists and knowledge representation practitioners was triggered, encouraging us to maintain and evolve the conversation. We turn this time towards the robotics community and aim at running the third edition of AnSWeR at IROS2019 to strengthen the dialogue and engage a new, wider audience, bridging the gap between these overlapping, but complementary communities. We also firmly believe that by hosting AnSWeR at IROS2019, we will be able to foster conversations among the Asian community, that put profound efforts into the robotics research, and the EU audience, which presents a long-standing tradition in knowledge management and representation, and the Internet of Things.

Topics of interest

Authors will be encouraged, but not limited, to consider the following set of topics:

  • Usability of available Semantic Web resources in Robotics
  • Semantic methods to support the development of robotic systems
  • Knowledge Representation and Reasoning techniques for Robotics
  • Knowledge-based systems for robots
  • Semantic solutions to enable spatiotemporal planning and reasoning
  • Ontologies and standardization of terminology for robotics applications
  • Semantic Mapping
  • Knowledge acquisition in robotic applications
  • Integration of local robotic knowledge with data from the Web
  • Planning and navigation using knowledge graphs
  • Robotics within the Web of Things/Internet of Cloud Robotic Things
  • Knowledge and perception
  • Semantic technologies to support Cloud Robotics systems
  • Semantic approaches for entity linking, grounding, and anchoring
  • Concrete use cases of working robotic systems exploiting semantic technologies
  • Future trends at the intersection of Robotics and Knowledge Representation
  • Knowledge Graphs, Linked Data and semantic technologies for robots
  • Impact and relation of knowledge-based technologies in social robotics and robot ethics
  • F.A.I.R. data for robotics systems
  • Knowledge-based embodied conversational agents (e.g., chatbots)

Challenge Task: Knowledge Extraction from robotic ontologies - Extracting the most popular terms

No idea yet to submit a research paper?

The workshop also offers a challenge task idea for doing analytics on LOV4IoT-Robotics ontology catalog following the research ideas from the Knowledge Extraction for the Web of Things Challenge (KE4WoT). Research can submit their research paper following the challenge idea.

Definition: Loading a set of robotic ontologies and extract the most popular/important terms (e.g., concepts and properties). It would be interesting to have algorithms answering such questions: (1) What are the sensors designed within the ontology (e.g. Body Thermometer)?, (2) What are the logical rules (IF THEN ELSE) designed within the ontology (e.g., if body temperature greater than 38 Degree Celcius than fever)? What is the applicative domain within this ontology (e.g., healthcare) useful when the ontology covers several domains (e.g., robots for cooking, robots for surgery).

Input: A set of ontologies from the LOV4IoT ontology catalog (e.g., robotic ontologies).

 LOV4IoT Tutorial to get robotic ontologies: http://lov4iot.appspot.com/?p=queryRoboticOntologiesWS (using a web service or a dump of ontologies)

Output: For each ontology, finding the most 10 relevant concepts and properties.

 Suggestion:
 Paper: Concept Extraction from Web of Things Knowledge Bases [Noura et al. 2018]
 Paper: OntoKhoj: a semantic web portal for ontology searching, ranking and classification [Patel et al. 2003] 
 Paper: Identifying potentially important concepts and relations in an ontology [Wu et al. 2008] 

Impact: Such algorithms would demonstrate the most relevant concepts and properties in a set of domains. Hopefully, the algorithm will be generic enough to be applied to any domains. Such algorithms could be relevant to assist to create iot.schema.org for instance.

Audience: Robotics communities who want to discover and study already designed models, any developers and/or data scientists willing to make statistics, Knowledge Extraction Experts.

Evaluation: For evaluation purpose, we will choose some ontologies referenced within the set of robotic ontologies mentioned above.

  Criteria 1: What are the sensors/devices described within the ontologies?
  Criteria 2: What are the concepts and properties which can be linked to other ontologies (e.g., Cooking concepts could be linked to food and recipe ontologies)?
 
  Check here our evaluation tables to show the most common concepts within robotic ontologies

Genericity: To test the genericity of your algorithm, you can play with additional datasets:

  LOV4IoT Tutorial to get food ontologies: http://lov4iot.appspot.com/?p=queryFoodOntologiesWS (using a web service or a dump of ontologies)

Important Dates

Workshop Paper Submission (Extended deadline): 15 August 2019

All deadlines are 23:59 Pacific Standard Time (PST)

Notification to authors (to confirm): 7 September 2019

Camera Ready of authors’ papers (to confirm): 15 September 2019 5:00 PM PDT

Workshop Date: 4 November 2019

Where: The Venetian Macao, Macau, China

Co-located with IROS 2019 (IEEE/RSJ International Conference on Intelligent Robots and Systems): https://www.iros2019.org/

Registration: https://www.iros2019.org/registration

Do you need a visa? https://www.iros2019.org/registration%26travel-visainformation  

Submission

Submission Web page: https://easychair.org/conferences/?conf=answer19

Papers must follow the IROS conference format: http://ras.papercept.net/conferences/support/support.php

Long research papers (up to 12 pages).

Short papers, position papers or demos (up to 6 pages).

The page limit includes references, and no other text whatsoever.

Feel free to ask any questions to answer19@easychair.org

Assuming a sufficient number of submissions, accepted contributions will be published as online proceedings courtesy of CEUR-WS. Moreover, best papers will be given the chance to submit an extended version of the work in a special issue of the Semantic Web Journal (more details TBA).

Program

Memento IROS regisration: https://www.iros2019.org/registration

ONGOING program here (or seen screenshot below)

Do you plan to attend our workshop? Feel free to present yourself in our Lightning Talks (1-minute slide) for attendees to ease networking and discussions or share any comments.

ProgramPage1 IROS2019AnswerWorkshop.png


Lars Kunze, Oxford University, UK Yuke Zhu, Stanford, USA Mathieu d'Aquin, Insight Center for Data Analytics, Ireland Todor Stoyanov, Orebro University, Sweden Veera Ragavan, Monash University, Malaysia Filippo Cavallo, The BioRobotics Institute Scuola Superiore Sant'Anna, Italy


Keynote 1: "Autonomous Robots in a Connected World"

Dr. Lars Kunze, Oxford University, UK (Already confirmed).

Dr. Kunze has a long-standing experience with knowledge representation formalisms for robotics and will present his work held in the context of the RoboEarth (http://roboearth.ethz.ch) and ALOOF (https://project.inria.fr/aloof/) projects.


Keynote 2: "Learning How-To Knowledge from the Web"

Dr. Yuke Zhu, Stanford, USA (Already confirmed)

Abstract: Recent advances in artificial intelligence have led to a series of successful applications, such as image recognition and machine translation. Along with great progress in machine learning models, web data has been serving a critical role as the training source to fuel these data-hungry models. However, large-scale web datasets, e.g., annotated image datasets and curated natural language corpora, have mostly focused on capturing the declarative knowledge ("what-is"), which stores the factual and descriptive information. While it has empowered AI algorithms to embrace a deeper understanding of our world, the interactive nature of robotics further requires a complementary type of knowledge: the procedural knowledge ("how-to") that determines how to interact with the world. In this talk, I will discuss two directions in learning how-to knowledge from the web. First, I will present our work on learning task structures from video data. We develop one-shot imitation learning methods to extract high-level task procedures from video demonstrations. Second, I will introduce our ongoing effort in building RoboTurk, a large-scale remote crowdsourcing platform. RoboTurk enables a parallel pool of online users to together create the largest dataset of human demonstrations for robot skill learning.

Bio: Yuke Zhu is currently a senior research scientist at NVIDIA and a visiting scholar at Stanford. He will be joining UT-Austin as an Assistant Professor in Computer Science starting Fall 2020. He received his master's and Ph.D. degrees from Stanford. His Ph.D. thesis centers around closing the perception-action loop to make robot intelligence more generalized and applicable to less-controlled environments. His research lies at the intersection of robotics, machine learning, and computer vision. He develops computational methods of perception and control that give rise to intelligent robot behaviors. Through his work, he aspires to teach robots to understand and interact with the visual world around them. His expertise has gained attention from a variety of news outlets, leading tech institutions, and award organizations. His publications have won several awards and nominations, including the Best Conference Paper Award in ICRA 2019. His work has been covered by media, such as MIT Technology Review and Stanford News. In addition, he's had research collaborations with Snap Research, Allen Institute for Artificial Intelligence, and DeepMind Technologies.


Keynote 3: “Virtualized knowledge for robot understanding”

Prof. Mathieu d’Aquin, Insight Center for Data Analytics, Ireland (Already confirmed)

Recent advances in artificial intelligence have led to a series of successful applications, such as image recognition and machine translation. Along with great progress in machine learning models, web data has been serving a critical role as the training source to fuel these data-hungry models. However, large-scale web datasets, e.g., annotated image datasets and curated natural language corpora, have mostly focused on capturing the declarative knowledge ("what-is"), which stores the factual and descriptive information. While it has empowered AI algorithms to embrace a deeper understanding of our world, the interactive nature of robotics further requires a complementary type of knowledge: the procedural knowledge ("how-to") that determines how to interact with the world. In this talk, I will discuss two directions in learning how-to knowledge from the web. First, I will present our work on learning task structures from video data. We develop one-shot imitation learning methods to extract high-level task procedures from video demonstrations. Second, I will introduce our ongoing effort in building RoboTurk, a large-scale remote crowdsourcing platform. RoboTurk enables a parallel pool of online users to together create the largest dataset of human demonstrations for robot skill learning.


Keynote 4: "Semantic mapping for robots and by robots: the role of high-level information"

Dr. Todor Stoyanov, Orebro University, Sweden (Already confirmed)

Semantic maps have the potential of serving as a bridge between classical high-level AI and low-level robot exploration and motion generation. Despite substantial research efforts however, the problem formalism with regards to semantic mapping is still largely disjoint: that is, AI and robotics researchers have very different notions with respect to semantic mapping. In this talk, I will attempt to offer a glimpse at the different notions of semantics with respect to robot maps and argue that top-down and bottom-up approaches to the problem are on the verge of finding common ground.

Dr. Todor Stoyanov is an Associate professor with the Center for Applied Autonomous Sensor Systems at Örebro University. His primary research area is on the intersection between perception and manipulation, with a focus on robot learning techniques for mobile manipulation.


Keynote 5: “An overview of IEEE 1872.2 WG Autonomous Robotics Ontology Progress”, and “Towards an Ontology driven Design and Development process”

Prof. Veera Ragavan, Monash University, Malaysia (Confirmed)


Keynote 6: "Internet of Robotic Things (IoRT)" (Cancelled)

Dr. Filippo Cavallo, The BioRobotics Institute Scuola Superiore Sant'Anna, Italy

But you can look at: Internet of Robotic Things – Converging Sensing/Actuating, Hyperconnectivity, Artificial Intelligence and IoT Platforms [Vermesan et al. 2017] (Chapter 4, p 124)

or Workshop on Robotics for Healthy Living and Active Ageing at IEEE ROMAN 2019


Paper 1: ABOUT THE MAN-MACHINE VERBAL INTERACTIONS (FRENCH LANGUAGE). Pierre-André Buvet, Abdelhadi Rouam, and Bertrand Fache.


Paper 2: Hybrid Question Answering System based on Natural Language Processing and SPARQL Query. Mickael Rajosoa, Rim Hantach, Sarra Ben Abbes, and Philippe Calvez.


Paper 3: Auto-Perceptive Reinforcement Learning (APRiL). Rebecca Allday, Simon Hadfield, and Richard Bowden.

Workshop Chairs

Ilaria Tiddi, Vrije Universiteit of Amsterdam, The Netherlands Masoumeh Mansouri, Örebro University, Sweden Emanuele Bastianelli, Heriot-Watt University, UK Amelie Gyrard, Kno.e.sis, Wright State University, USA Institutes.png


Ilaria Tiddi

Homepage: http://kmitd.github.io/ilaria

Institute: Knowledge Representation and Reasoning group of the Vrije Universiteit of Amsterdam (NL)

Dr. Ilaria Tiddi is a Research Associate in the Knowledge Representation and Reasoning group of the Vrije Universiteit of Amsterdam (NL). Ilaria's research focuses on responsible AI and, in particular, she is working on the creation of transparent/explainable systems using knowledge from large, heterogeneous knowledge graphs. Ilaria was involved in a number of research activities, e.g. the organization of the Knowledge Capture conference (of which she is also Steering Committee member), the CEUR-WS Editorial Board, the 2 editions of the Linked Data 4 Knowledge Discovery workshop (LD4KD2014, LD4KD2015). She also assisted the organization of the Semantic Web Summer Schools in 2015 and 2016.

Masoumeh (Iran) Mansouri

Homepage: https://iranmansoori.github.io/

Institute: Cognitive Robotic Systems of the AASS center at the Örebro University, Sweden / Birmingham University, UK

Dr. Masoumeh (Iran) Mansouri is a Research Associate at the Cognitive Robotic Systems of the AASS center at the Örebro University, where she also received her Ph.D. (2016). Her research is primarily related to Knowledge Representation and Reasoning (KR&R) for Robotics. She has focused on hybrid methods that integrate automated task and motion planning, scheduling, as well as temporal and spatial reasoning. Her overall passion is to work towards the realization of fully AI-driven integrated robotic systems. Iran has participated in a number of Robotics-related research activities, including the organization of the IROS2016 workshop on Integrating Multiple Knowledge Representation and Reasoning Techniques in Robotics (MIRROR-16); a lecturer of several editions of the International Winter Schools on “Artificial Intelligence and Robotics” (LUCIA Winter School); a tutorial instructor on Integrated Motion Planning, Coordination and Control for Robot Fleets at ICAPS 2018; and an invited speaker for the cognitive robotics workshop at ICRA 2018.


Emanuele Bastianelli

Homepage: https://www.hw.ac.uk/staff/uk/macs/Emanuele-Bastianelli.htm

Institute: Heriot-Watt University, Edinburgh, UK

Dr. Emanuele Bastianelli is a Research Associate at the Heriot-Watt University, Edinburgh, UK. He got his Ph.D. in Computer Science from the University of Rome Tor Vergata. He previously worked as a Research Associate at the Knowledge Media Institute of the Open University, and as a Research Assistant at the Cognitive Cooperating Robots Lab (Lab Ro.Co.Co.) of Sapienza University of Rome. His research mostly focuses on Machine Learning applied to Natural Language Processing, with a specific target on Language Understanding for robotic applications and Conversational AI. During his career, he covered also other areas such as Information Retrieval and Urban Data Mining. Emanuele was the main organizer of the first AnSWeR workshop held at ESWC2017.


Amelie Gyrard

Homepage: http://sensormeasurement.appspot.com/?p=AmelieGyrard

Institute: Kno.e.sis, Wright State University, USA

Dr. Amelie Gyrard (http://wiki.knoesis.org/index.php/AmelieGyrard) is a post-doc researcher with the appointment as a research assistant professor at Kno.e.sis, Wright State University, USA. Previously, she worked at MINES Saint-Etienne (Connected Intelligence - Knowledge Representation and Reasoning team), France. She was also a post-doc at Insight Center for Data Analytics, National University of Galway, Ireland and actively worked in the scientific development and coordination of the FIESTA-IoT (Federated Interoperable Semantic IoT/Cloud Testbeds and Applications) EU H2020 project. Her research interests are Software engineering for Semantic Web of Things and Internet of Things (IoT), semantic web best practices and methodologies, ontology engineering, Artificial Intelligence (AI) such as reasoning, and interoperability of IoT data. She received her Ph.D. in 2015 from Eurecom where she designed and implemented the Machine-to-Machine Measurement (M3) framework. One of its components, called PerfectO, aims to disseminate and facilitate Ontology Best Practices. She is also a reviewer for IoT, Semantic Web related journals, and conferences. Dr. Amelie Gyrard has been co-organizer of the following events: 9th International Semantic Sensor Networks Workshop (SSN) 2018 co-located with International Semantic Web Conference (ISWC) 2018, SmartIoT 2018 Workshop: AI enhanced IoT data processing for Intelligent Applications at AAAI 2018, Semantic Web meets Internet of Things and Web of Things Tutorial [3rd Edition] co-located with ISWC 2017, WWW 2017, and ISWC 2016.

Programme Committee (PC)

List of technical program committee members already agreed:

  • Sandro Fiorini, IBM, Brazil
  • Valerio Basile, University of Turin, Italy
  • Marc Hanheide, University of Lincoln, UK
  • Marcos Barreto, Universidade Federal da Bahia, Brazil
  • Paulo Goncalves, IDMEC, Instituto Politecnico de Castelo Branco, Portugal
  • Partha Pratim Ray, Sikkim University, India
  • Edison Pignaton De Freitas, Universidade Federal do Rio Grande do Sul, Brazil
  • Stefano Borgo, Institute of Cognitive Sciences and Technologies (ISTC), Italy
  • Marjan Alirezaie, Orebro University, Sweden
  • Daniela D'Auria, University of Naples Federico II, Italy
  • Daniel de Leng, Linköping University, Sweden
  • Sonia Bilbao, TECNALIA, Parque Tecnologico de Bizkaia, Spain
  • Filippo Cavallo, Sant'Anna, Italy
  • Mohammed Diab, University of Catalona, Spain
  • Veera Ragavan Sampath Kuma, Monash University, Australia
  • Davide Bacciu, Universita di Pisa, Italy
  • Hirenkumar Chandrakant Nakawala, University of Verona, Italy
  • Joel Luis Carbonera, Federal University of Rio Grande do Sul, Brazil
  • Andrea Orlandini, Institute of Cognitive Science and Technology, National Research Council of Italy, Italy
  • Joanna Olszewska, University of the West of Scotland, UK
  • Julita Bermejo-Alonso, Universidad Politecnica de Madrid, Spain
  • Maki Habib, The American University in Cairo, Egypt
  • Enrico Daga, Knowledge Media Insitute of The Open University, UK
  • Martin Günther, Robotics Innovation Center at German Research Center for Artificial Intelligence (DFKI), Osnabrück Branch, Germany
  • Claudio Gallicchio, University of Pisa, Italy
  • Stefan Schiffer, RWTH Aachen University, Germany
  • Alessandro Russo, Drim Robotics Sp. z.o.o., Poland
  • Joao Quintas, Institute for Systems and Robotics, Lisboa, Portugal
  • Martin Serrano, Insight Center for Data Analytics, Ireland

Past Workshop Editions and Related Workshops

  • MIRROR workshop at IROS 2016. MIRROR-16, held at IROS in 2016, focused on issues related to the integration of several reasoning systems., i.e. the problem of jointly reasoning about heterogeneous and interdependent aspects of the world, expressed in different forms and at different levels of abstraction.
  • AGAINST workshop at ICRA 2019. AGAINST workshop focuses on bias-sensitization of robot behavior on issues such as gender, race, age, and culture. The goal of this workshop is to raise awareness of and provide new insights into, the biases involved in robot models, human-robot interaction, datasets, and knowledge representation; as well as the techniques which will be required to avoid harmful bias and discrimination by robots.

References