News: I HAVE MOVED!!! (as of July 2007: contact, position)

Benjamin Grosof, PhD
Senior Research Program Manager, Knowledge Systems, at Vulcan Inc.
Principal, Benjamin Grosof & Associates LLC (part time consulting practice).
Previously an MIT Sloan professor, IBM Research scientist, and RuleML co-founder.

Expertise in IT Research -- Science, Design, and Leadership:
Semantic Technology, Semantic Web, Info/Knowledge Integration, and AI
for E-Business, E-Services, and E-Science; particularly:
Rules and Business Policies; Semantic Social Networking and Web 3.0; E-Commerce, Financial, and Trust.

Note that the rest of my website's content has not yet been updated to reflect the move to Vulcan, except for the contact info and position description. Check back in a few weeks...

My webpage has also moved! It's now http://www.mit.edu/~bgrosof/ rather than http://ebusiness.mit.edu/bgrosof/.

 
From July 2000 until the end of June 2007:

Benjamin Grosof
Assistant Professor in Information Technology,
MIT Sloan School of Management

"Creating and Studying Knowledge-based Web Technologies for E-Commerce"


Contact   
Bio     
Papers, Talks, Software    *News: SweetRules Released*
Research Projects Overview     *News: Tutorial Available*
RulesKR Project:  Rules Representation for Semantic Web Services   
BizSWS Project:  Business Implications of Semantic Web Services   
Semantic Web Services Primer, SWSI   
XML Rules Standards,  RuleML, SWRL   
Groups at MIT Sloan   
Teaching   
Professional Service Activities   
Consulting   
Upcoming/Recent Activities    *News: W3C Rules Working Group*
Search   

 

Research Projects Overview:

  1. RulesKR: Rules Knowledge Representation Technologies, esp. for Semantic Web Services
    • RuleML Logic Programs + Ontologies + Databases

  2. BizSWS: Business Implications of Semantic Web Services, including Applications and Strategy
    • E-Contracting, Business Policies: ... in B2B, Supply Chain, Finance
    • More generally: Knowledge Integration and Business Process Automation; Business Rules, Intelligent Agents, Business Intelligence, Knowledge-based E-Markets

 

RulesKR Project on Rules Knowledge Representation Technologies, especially for Semantic Web Services

This project is to create and study fundamental technologies for rules knowledge representation, for infrastructural use in Semantic Web Services. It includes fundamental reasoning theory (including extensions to logic programs), technology design (e.g., architecture, algorithms) and prototypes, and standards proposals (including RuleML). The project has created a highly capable integrated toolkit called SweetRules, available in open source. ("Sweet" stands for "Semantic WEb Enabling Technologies".) In addition, there is a close dialectic with exploring applications scenarios (drawn from our project on business implications), and strategies.

This project is concerned largely with communication of rule-form beliefs (information), assimilation of such beliefs from multiple sources, reasoning about the scope and degree of trust of those sources, handling of conflicts between those sources, and inter-operable executability of inferencing with those beliefs via knowledge-based and database systems. The sources might be agents, applications, or databases, for example. The focus is especially on information about business rules or policies, including in e-contracts. The technical approach is based on declarative logic programs.

Topics include:

  • rules inter-operability and translation among heterogeneous rule systems, especially currently commercially important (CCI) families of rule systems including:
    • relational database systems (SQL)
    • event-condition-action (ECA) rule systems
    • production rule systems (whose ancestors include OPS5, CLIPS, Jess)
    • Prolog
  • rules standardization, especially about RuleML and its relationship to other emerging Web standards including OWL from W3C WebOnt (Semantic Web-Ontologies) Working Group. I co-lead the RuleML standards effort.
  • rules expressiveness: (which involves plenty of theory, e.g., about non-monotonic reasoning or computational complexity)
    • prioritized conflict handling, merging, updating ("courteous logic programs")
    • procedural attachments, linkages to general business processes and procedural code ("situated courteous logic programs")
    • use of description logic ontologies cf. W3C's OWL ("description logic programs")
    • use-of / integration-with other general-purpose W3C Web standards, e.g., RDF, URI, and perhaps XInclude.
  • rules inferencing, including triggered actions, both forward-direction (data-driven or event-driven) and backward-direction (query-answering).
  • merging, assimilation, and updating: of rules knowledge from multiple sources. I.e., automated learning by "talking" as well as from data mining.
  • rule-based semantic web services, including architectural and application roles of rules in semantic web services
  • overall, knowledge integration. (NB: sometimes a.k.a. "intelligent information integration".)

To achieve practical e-commerce applications, often one must also strive in the rules KR design to enable:

  • software engineering "-ilities", e.g., embeddability, modularity, scaleability, integration architectures
  • knowledge representation and inferencing about probabilities

SWEET:
SWEET ("Semantic WEb Enabling Technology") is an overall set of tools I (with collaborators) am developing. It includes SweetRules (supporting Situated Courteous Logic Programs in RuleML) and SweetDeal (supporting e-contracting; see below). Prototypes of SweetRules and SweetDeal have been running since late 2001. For more about Sweet, see the description of the software.

DAML:
This project is receiving significant funding support from the DARPA Agent Markup Language (DAML) Program, whose overall purpose is to develop techniques for high-level communication between agents in XML. I am Principal Investigator (PI) at MIT for this DAML grant award. SWSI grew in part out of the DAML program.

I co-lead (with Mike Dean) the DAML Rules effort and the Joint Committee Rules effort, which are closely related to RuleML.

Tim Berners-Lee, inventor of the Web, and head of the World Wide Web Consortium, is PI at MIT of another DAML grant project about the Semantic Web. I work with Tim and the MIT LCS / W3C team he leads, as well as with several other researchers outside of MIT, on DAML and the Semantic Web.

You can see my recent papers and talks for more about the RulesKR project. See also SWSI especially its SWSL language effort which includes rules.

 

BizSWS Project on Business Implications of Semantic Web Services, including Applications and Strategy

(NB: "BizSWS" is pronounced "BizSwizz")

This project is to create and study the business implications of Semantic Web Services (SWS). It includes applications design and scenarios, analysis of business value, strategy, and theory. The applications largely focus on using rules for e-contracting, web services, and financial knowledge integration.

SWS offers the promise of dramatically increasing the degree of automation (and lowering costs) in machine-to-machine/application-to-application communications and business processes, as compared to the first generation of the Web which is primarily oriented towards human-to-machine/human-to-application interactions.

The fundamental rules technologies in our RulesKR project above are motivated by SWS e-commerce applications and strategies, e.g., for e-contracting and finance, and are developed in tandem with them. Developing the fundamental technologies in tandem with the applications provides bi-directional feedback. This helps focus both sides. The e-contracting prototype is called SweetDeal (more about that below).

Topics include:

  • e-contracting (e.g., in our SweetDeal approach):
    • representing deal descriptions (offerings, bids), and particularly their contingent provisions (e.g., handling late deliveries, refunds, or payment problems)
    • descriptions of goods (products or services) and their pricing
    • negotiating proposed deals, to create contracts
      • generating proposals; inferencing about proposals to evaluate them; modifying proposals; auctions
    • matchmaking of business partners, in the discovery phase of contracting; sourcing; targeted advertising
    • executing, and monitoring, the performance of contracts
    • exploiting business process descriptions
  • particularly, deals about Web Services and e-services
    • focus on the "deal layer" of web services (or e-services), i.e., what the service does, at what price, why one would want to purchase it, what guarantees are offered, and what are the terms & conditions of the deal -- as opposed to the lower "mechanics layers" for how to invoke or register a web services (e.g., WSDL) which have been to date the main focus of industry standardization efforts
  • integration of financial information/knowledge from multiple sources (e.g., in our ECOIN approach)
  • risk management in e-contracting and finance
  • business policies, more generally, e.g., in:
    • pricing
    • customer relationship management (CRM) including customer service, promotions, targeting
    • trust/authorization, e.g., for security
    • privacy, e.g., in the W3C P3P standards effort
  • e-markets composed of knowledge-based applications ("agents")
  • virtual organizations via highly automated outsourcing of services
  • applications and strategy, more generally
    • early adopter industries and players
    • business value
    • sequencing of adoption and development; accelerants and catalysts
    • standards; role of open source and freeware

Semantic Web Services Initiative (SWSI) (pronounced "swizzie") coordinates and performs SWS research and early standards activities.

The Center for eBusiness @ MIT:
This project has in past received significant funding support from the Center for eBusiness @ MIT, a very large research center with dozens of company sponsors.

SweetDeal:
The e-contracting applications design, prototype, and scenarios, including for deals about e-services/web-services, are together called "SweetDeal". ("Sweet" stands for "Semantic WEb Enabling Technology".)

Extended COntext INterchange (ECOIN) is an approach to information integration involving mapping between different contexts of information usage or information supply. Those different contexts utilize different ontologies. Knowledge-based techniques for mapping between these heterogeneous ontological contexts then can create considerable value in financial applications.

You can see my recent papers and talks for more about the Biz+SWS project; there are (currently) three papers about e-contracting / SweetDeal, and two about financial knowledge integration / ECOIN. See also SWSI, especially its application scenario and industrial partnership aspects.

History:
This Biz+SWS project grows in part out of my previous work at IBM Research during 1994-2000, which was on:

    • Business rules for e-commerce. This resulted in IBM CommonRules which pioneered rules inter-operability and conflict handling, using the technical approach based largely on declarative logic programs in Java and XML which has been continued in SWEET and RuleML. An earlier version of this technology was IBM Agent Building Environment which was piloted in applications for personalized information workflow.

    • Intelligent agents and e-commerce/supply-chain applications that use business rules. This included being a Principal Investigator for a $29 Million industry-government consortium project called EECOMS (half-funded by NIST) during 1998-2000 on supply chain collaboration in agile manufacturing which used CommonRules and courteous logic programs in XML for application scenarios in e-contracting (procurement, supply chain, negotiation, exception management). Other applications included in catalogs & storefronts, security/authorization/trust, and personalization.

    You can see my (old) IBM project page there. The IBM CommonRules project continues under the leadership of Hoi Chan.

 

Intro and Primer: Semantic Web, Web Services, E-Commerce, Semantic Web Services

My research overall is concerned with the design and management of how automated enterprises and intelligent agents will soon communicate at a high level of shared understanding ("semantics") with each other over the Web in e-commerce (esp. B2B). Two important technical aspects of this are (1.) XML and (2.) techniques for knowledge representation and inferencing, especially for rules and ontologies.

An "ontology" is a formally specified set of vocabulary definitions. A "rule" is an if-then implication. Rules mention relations and other logical constants, and thus can rely on ontological definitions of those. "Knowledge representation" (KR) means what form of knowledge can be expressed, including both syntactic encoding and underlying semantics of meaning. This semantics is defined in terms of what conclusions are sanctioned from a given set of premises (e.g., rules or ontological definitions) in a particular chosen KR. The semantic aspect of KR is important to enable an agent/application to anticipate what another agent/application will believe/draw from a given set of communicated statements (i.e., exchanged information/knowledge).

My work is thus closely related to several aspects of the Semantic Web, an overall concept for the next generation of the Web, in which the Web becomes a repository of information that is automatically readable by programs in a way that has substantial semantics (i.e., becomes "agent-enabled"), rather than only human-readable/understandable as in the first generation of the Web. For more about the Semantic Web, you can see the W3C Semantic Web Activity and a Semantic Web community portal.

Within the Semantic Web overall, I am especially focusing on

  • rules knowledge representation: fundamental reasoning theory (including extensions to logic programs), technology (including SweetRules), and standards (including RuleML which I co-chair); and
  • business applications and strategy, e.g., e-contracting (including SweetDeal), intelligent knowledge integration (including ECOIN for finance), business policies (including authorization), and (more generally) business intelligence.

Another emerging concept is Web Services -- the delivery of electronic services using Web protocols. These services might be provided by invoking almost any kind of program, so this is an extremely broad concept. Part of my work is concerned with infrastructural services for the Semantic Web, e.g., for relatively broad-purpose knowledge translation and inferencing. Another part of my work is concerned with application-specific services that make use of the Semantic Web, e.g., services for e-contracting or financial knowledge integration. For more about Web Services, you can see the W3C Web Services activity and the Web Services Interoperability (WS-I) organization.

"Semantic Web Services" (SWS) is the convergence of Web Services and Semantic Web. SWS is the next major generation of the Web, in which e-services and business communication become more knowledge-based and agent-based. SWS includes both the infrastructural and the application-specific services I described above. It can be parsed as "{Semantic Web} Services" or as "Semantic {Web Services}". Until 2004, Semantic Web and Web Services were largely decoupled in industry standards and development efforts. However, since mid-2002, a research community with aspiration towards standards has formed around SWS, especially in the US and Europe. For overviews of Semantic Web Services, see my recent tutorials and the talks that mention it in their title. You can see a quickie 2003 list of Web resources about SWS, the Resources pages at SWSI, and the 2003 Panel on SWS at the WWW-20003 Conference.

Semantic Web Services Initiative (SWSI) (pronounced "swizzie") coordinates and performs SWS research and early standards activities.

 

XML Rules -- Standards; RuleML:

I am Co-Founder and Co-Chair of RuleML, an early-phase standards effort on a markup language for rules in XML, and more generally for Rules as part of the Semantic Web. The goal of this RuleML Initiative is eventual adoption as a Web standard, e.g., via the World Wide Web Consortium (W3C).

Research aspect: Along the way there are a number of interesting new research issues. Quite a bit of new theory and design work is required to support highly-expressive inferencing and interoperability in RuleML. Most of my recent publications and talks (since early 2001) are in part about extending or applying RuleML. Much of my recent research (with collaborators) is embodied in the:

  • New & Cool SweetRules V2. Created by a multi-institutional effort led by me, SweetRules is a powerful open source software integrated toolkit for semantic web rules, released Dec. 2004, which revolves largely around RuleML and provides a set of interoperable reference implementations for RuleML.
Also, RuleML largely grows out of Business Rules Markup Language (BRML) which I developed in my previous work at IBM Research and which is implemented in IBM CommonRules.

Scientific Conference annually on RuleML-related research topics: RuleML-2005, the first annual International Conference on Rules and Rule Markup Languages for the Semantic Web, was held in conjunction with the fourth International Semantic Web Conference (ISWC), in Nov. 2005. This is successor to three previous annual Workshops on the same topic, held in conjunction with the first three ISWC's. I co-founded the workshop series and conference, and was general co-chair for RuleML-2005. The RuleML-2006 conference will be held in Nov. 2006, in conjunction with ISWC-2006 in Athens, Georgia, USA.

The RuleML Initiative: The RuleML Initiative began in fall 2000, and released a first public version of XML DTD's for several rule flavors in January 2001. My co-founder/co-chairs are Harold Boley and Said Tabet. There are now several dozen participating institutions in the RuleML Initiative (a mix of industry and academe), and over a dozen prototype RuleML tools already available (supporting rule translation, inferencing, or authoring). Weekly group telecons and emails discuss technical as well as organizational issues.

Website(s): The main RuleML website is in process of being radically redesigned. Also, at any given time, partly due to lags in updating, there is also very salient stuff HERE about RuleML NOT on that site. E.g., see my recent papers and talks.

Rapid Growth in Influence: RuleML has grown rapidly in influence since its inception. Notable events since spring 2002 include that:

  • The DAML program has been (as of approx. Aug. 2002) using RuleML as its main starting point/focus for the DAML Rules effort (which I co-chair with Mike Dean), motivated also by having identified a requirement to use/support rules in the DAML-Services effort and its successor SWSI.

  • In particular, Rules have been (as of approx. Aug. 2002) the main focus of the Joint US/EU ad hoc Agent Markup Language Committee (usually called simply the "Joint Committee") which produced DAML+OIL, upon which the W3C OWL web ontology language is closely based. RuleML is the main starting point/focus for the Joint Committee's Rules effort, which I co-lead along with Mike Dean. The Joint Committee's email archive is a good place to find recent technical discussions on RuleML. (A RuleML email archive for organizational discussions is in process of being set up, as well.)

  • Tim Berners-Lee has disseminated (as of approx. Apr. 2002) an updated Semantic Web stack diagram (in png; or get pdf version). That proposes to include Rules as "on deck" (i.e., next) after Ontologies as an area for standardization within the Semantic Web. The W3C Semantic Web Activity's charter includes rules as an important area to standardize soon.

  • The Semantic Web Services Initiative (SWSI), the leading organization for developing early-phase standards proposals in semantic web services, has adopted RuleML as its rule markup language, and is collaborating on design of extensions and refinements to RuleML.

  • The Object Management Group, an important umbrella standards body for interoperability, is using RuleML as the approach to markup in its production rules standardization working committee.

  • RuleML (FOL and SWRL-rules sublanguages) is an acknowledged Member Submission to the World Wide Web Consortium (W3C).

  • The W3C is planning a Workshop in late spring 2005 on requirements and approaches to standardizing web-interoperable rules, drawing largely on the groundwork laid by RuleML. This is typically the penultimate major step before forming a W3C Working Group full-blown standards effort.

  • A Google search on "RuleML" yields 20,000+ hits (as of Feb. 2005).

 

Group Contexts:

  • MIT Sloan IT Group and associated Research Centers:
    My closer colleagues in the Sloan Information Technology (IT) group include faculty Erik Brynjolfsson, Stuart Madnick, Tom Malone, Wanda Orlikowski, Jack Rockart (mostly retired), and Peter Weill, and principal research scientists Mark Klein and Michael Siegel. An IT group web page (in need of updating) is available.

    I am member of two research centers and a large research project with which several of these colleagues are also affiliated:

    • Center for eBusiness@MIT (CeB), directed by Erik Brynjolfsson. CeB is a very large and vigorous center with dozens of company sponsors. It is centered at Sloan, involves many faculty outside of IT, and includes non-Sloan MIT faculty as well. Work in CeB includes several aspects of my research on e-contracting and semantic web services, which CeB has supported through its Vision Fund.
    • Center for Coordination Science (CCS), led by Tom Malone. CCS is the home of the Process Handbook (PH) project on business process design and knowledge management, which is integral in several aspects of my research.
    • Context Interchange (COIN) Project, led by Stuart Madnick. I contribute to the COIN project on use of information integration including financial reporting and services, using rule and ontology techniques.
    • Another research center, closely associated with the IT group, in which I participate as yet informally, is: the Center for Information Systems Research (CISR).

  • DAML; Tim Berners-Lee, W3C, MIT Computer Science & Artificial Intelligence Lab (CSAIL):
    I collaborate extensively with other researchers in the DAML Program, including the MIT CSAIL / W3C team led by Tim Berners-Lee, especially on semantic web rule technologies and standards, and their uses in policies and services.

 

Papers, Invited Talks, Software, etc. (selected)

      Preface Notes:
      • "Recent" here means since approximately summer 1999; "Earlier" means before then.
      • "Papers" here means: journal, conference/workshop, and working papers; and technical reports as well.
      • Recent papers are organized to group together successor/predecessor versions, as papers evolve to be extended and revised. I.e., each entry also often includes an extended/revised working paper version and/or earlier versions.
        • Conference/workshop talk slides, too, are usually included with the papers.
        • For chronological sequencing purposes, the date is taken to be that of the most recent refereed publication version or, lacking that, of the most recent working paper or technical report version.

 

Selected Recent Papers Categorized by Topic: (chronological within each category)

!!NOTE!!: the MOST RECENT papers (e.g., last several months) may not yet be listed here, but ONLY in the CHRONOLOGICAL listing of papers, due to lags in website maintenance/editing.

 

*RulesKR: Rules and Ontologies knowledge representation for the Semantic Web and Semantic Web Services:

    • Including: Situated and Courteous Logic Programs; Description Logic Programs; RuleML, SWRL, SWSL, and rule system Interoperability; and Courteous Inheritance. (With motivating examples usually from e-commerce policies and processes.)

*BizSWS: Business Implications of Semantic Web Services:

*Other Knowledge-Based E-Commerce (including AI, agents):

 

Selected Earlier Papers Categorized by Topic: Refereed Publications and Research Reports (1995-1999)

*Courteous Logic Programs (Earlier):

*Examples and demos of business rules using Courteous LP for e-commerce (Earlier):

*Situated Logic Programs (Earlier):

*XML Agent Communication (Earlier):

 

Selected Recent Papers organized Chronologically: (but grouped so that successor/predecessor versions are together)

 

Recent Standards Proposal Reports:

  • "Rule Markup Language Initiative (RuleML)". 2001-present

    • Standards Proposal Reports and Material on the RuleML language and Initiative. website.
      Currently Versions 0.8+ of 2004-2005, revised from Versions 0.7+ of Jan. 2001 - 2003.
      By Harold Boley, Benjamin Grosof, Said Tabet, and additional collaborators in the RuleML Initiative (NB: authorship among these three is alphabetic). Includes extensive documentation, news, discussion, summaries, presentations.

      Comment: The leading emerging standard for interoperable web rules in XML, including for semantic web and business rules. It is based on declarative logic programs, including situated courteous logic programs (SCLP), along with first order logic.
      The ECRA journal paper, and the ISWC-2006 Tutorial slides, each provide an overview of RuleML including particularly SCLP RuleML.
      SweetRules V2 provides a set of reference implementations in open source for RuleML inferencing and translation, supporting SCLP and SWRL.

      RuleML is being used heavily by the Semantic Web Services Initiative (SWSI) in its Language (SWSL), and is beginning to be used by Object Management Group (OMG) in its production rules standards committee. (As of Feb. 2005.)

      A Google search on "RuleML" yields a hit count of over 20,000 as of Feb. 15, 2005.

      Revisions and extensions are in progress, with a new major version planned for release in spring or summer 2005.

  • "FOL RuleML: The First-Order Logic Web Language". 2004-present

    • Standards Proposal Research Report.
      Version 0.9 of 2004-11-02 (revised from version 0.7 of 2004-08-10.)
      By Harold Boley, Mike Dean, Benjamin Grosof, Michael Sintek, Bruce Spencer, Said Tabet, and Gerd Wagner. An emerging industry standards proposal that is an ...
      Acknowledged W3C Submission. A W3C Submission once acknowledged by W3C (the World Wide Web Consortium) becomes a technical report document of W3C.

      Comment: An interoperable web syntax for First Order Logic, as an addition to the RuleML family of sublanguages.

  • "SWRL: A Semantic Web Rules Language Combining OWL and RuleML". Report, 2004.

  • "Semantic Web Services Language Requirements". 2003-present

    • Version 1 of Oct. 2003.
      Edited by Benjamin Grosof, Michael Gruninger, Michael Kifer, David Martin, Deborah McGuinness, Bijan Barsia, and Austin Tate (NB: editorship order is alphabetic). The editors are also the primary authors, but the full set of authors includes a larger set of members of the SWSL Committee. SWSL is the Language part of the Semantic Web Services Initiatve (SWSI).

      Comment: Revised version is in progress, in tandem with preparation of major SWSL design research report planned for release in spring 2005.

  • "Semantic Web Services Language". In Preparation 2004-present

    • Version 1 preliminary draft, planned for release in spring 2005.
      By Steve Battle, Daniela Berardi, Benjamin Grosof, Michael Gruninger, Rick Hull, Michael Kifer, David Martin, Sheila McIlraith, Jianwen Su, and others. (NB: authorship set is preliminary and its order is alphabetic). These are the members of the SWSL Committee. SWSL is the Language part of the Semantic Web Services Initiatve (SWSI).

 

Invited Talks: (slidesets -- usually quite detailed;   includes tutorials)