Knowledge graphs have gained popularity for their ability to organize and analyze complex data effectively. When combined with graph embedding techniques, such as graph neural networks (GNNs), knowledge graphs become a potent tool in providing valuable insights. This study explores the application of graph embedding in identifying competitors from a financial knowledge graph. Existing state-of-the-art(SOTA) models face challenges due to the unique attributes of our knowledge graph, including directed and undirected relationships, attributed nodes, and minimal annotated competitor connections. To address these challenges, we propose a novel graph embedding model, JPEC(JPMorgan Proximity Embedding for Competitor Detection), which utilizes graph neural network to learn from both first-order and second-order node proximity together with vital features for competitor retrieval. JPEC had outperformed most existing models in extensive experiments, showcasing its effectiveness in competitor retrieval.

Knowledge Graphs have emerged as a compelling abstraction for capturing key relationship among the entities of interest to enterprises and for integrating data from heterogeneous sources. JPMorgan Chase (JPMC) is leading this trend by leveraging knowledge graphs across the organization for multiple mission critical applications such as risk assessment, fraud detection, investment advice, etc. A core problem in leveraging a knowledge graph is to link mentions (e.g., company names) that are encountered in textual sources to entities in the knowledge graph. Although several techniques exist for entity linking, they are tuned for entities that exist in Wikipedia, and fail to generalize for the entities that are of interest to an enterprise. In this paper, we propose a novel end-to-end neural entity linking model (JEL) that uses minimal context information and a margin loss to generate entity embeddings, and a Wide & Deep Learning model to match character and semantic information respectively. We show that JEL achieves the state-of-the-art performance to link mentions of company names in financial news with entities in our knowledge graph. We report on our efforts to deploy this model in the company-wide system to generate alerts in response to financial news. The methodology used for JEL is directly applicable and usable by other enterprises who need entity linking solutions for data that are unique to their respective situations.

Cognitive simulation of analogical processing can be used to answer comparison questions such as: What are the similarities and/or differences between A and B, for concepts A and B in a knowledge base (KB). Previous attempts to use a general-purpose analogical reasoner to answer such questions revealed three major problems: (a) the system presented too much information in the answer, and the salient similarity or difference was not highlighted; (b) analogical inference found some incorrect differences; and (c) some expected similarities were not found. The cause of these problems was primarily a lack of a well-curated KB and, and secondarily, algorithmic deficiencies. In this paper, relying on a well-curated biology KB, we present a specific implementation of comparison questions inspired by a general model of analogical reasoning. We present numerous examples of answers produced by the system and empirical data on answer quality to illustrate that we have addressed many of the problems of the previous system.

Biological processes can be expressed from multiple perspectives. One basic perspective involves specifying the process steps, how they are ordered, their participants, and how participants are involved in different steps. Two additional perspectives on processes are how energy is transferred and how the processes are regulated. Such perspectives are essential for understanding the basic mechanisms of biology. We consider a representation of energy transfer and regulation for biological processes, and consider several example questions that can be answered using this representation. Our work is driven by the description of processes in an introductory biology textbook and provides the ontology design for an education application.

KB_Bio_101 contains knowledge about processes and mechanisms, and was created from an introductory textbook in biology. We give an overview of its content, summarize the key concepts represented, and give some examples of problems requiring further ontology research.

In the fall 2010 issue of the AI Magazine, we reported the design, implementation and evaluation of a knowledge acquisition system called AURA. AURA enables domain experts in Physics, Chemistry and Biology to author their knowledge, and a different set of experts to pose questions against that knowledge. The evaluation results previously reported were from 50 pages each from science textbooks in Physics, Chemistry and Biology. The results were most promising for Biology. Based on those results we undertook a content building effort to capture knowledge from approximately 315 pages (or 20 chapters) of the same Biology textbook [2] and incorporated the resulting content in the electronic version of that book. In this demo/poster session, we will demonstrate the biology knowledge base (KB) created using AURA, the electronic textbook application Inquire, and discuss the knowledge engineering process we used to construct the KB.

To scale the knowledge base of a Biology textbook from 50 pages to 300 pages in the context of Project Halo, we have formulated a knowledge factory process [1]. The process involves a sentence-based encoding strategy under which a domain expert examines each sentence in the text-book [2] and represents it in a knowledge base (KB) as best as it can be represented. While encoding each sentence, at least two hard problems must be addressed: defining what it means to represent a sentence and defining the necessary ontology primitives that should be used in that representation. The purpose of this poster presentation is to explain these problems in more detail, and discuss the operational solutions that we have adopted for solving them.

Our long-term goal is to build a query answering system that can answer questions on a wide variety of topics and explain the answers. In such a situation, a designer faces the challenge of how to specify the KR&R requirements that are needed to answer questions. In this paper, we introduce a categorization of KR&R methods, and apply it to specifying the requirements for answering questions in six different domains: Physics, Chemistry, Biology, Environmental Science, Microeconomics, and U.S. Government & Politics. Drawing from the corpus of about 500 questions that we analyzed, we consider an example question in each domain and show the analytical process that we used to derive the requirements in terms of the KR&R categorization. We analyze the effectiveness of the current KR&R categorization, and identify directions for future work suggesting how this categorization can be further evolved by community participation.

The long-term goal of Project Halo is to build an application called Digital Aristotle that can answer questions on a wide variety of science topics and provide user- and domain-appropriate explanations. As a near-term goal, we are focusing on enabling subject matter experts (SMEs) to construct declarative knowledge bases (KBs) from 50 pages of a science textbook in the domains of Physics, Chemistry, and Biology in a way that the system can answer questions similar to those in an Advanced Placement (AP) exam in the respective discipline. The textbook knowledge is a mixture of textual information, mathematical equations, tables, diagrams, and domain-specific representations such as chemical reactions. In this paper, we explore the following question: Can we build a knowledge capture system to enable SMEs to construct KBs from the knowledge found in science textbooks and use the resulting KB for deductive question answering? We answer this question in the context of a system called AURA that supports knowledge capture from science textbooks.

As part of the ongoing project, Project Halo, our goal is to build a system capable of answering questions posed by novice users to a formal knowledge base. In our current context, the knowledge base covers selected topics in physics, chemistry, and biology, and our question set consists of AP (advanced high-school) level examination questions. The task is challenging because the questions are linguistically complex and are often incomplete (assume unstated knowledge), and because the users do not have prior knowledge of the system's contents. Our solution involves two parts: a controlled language interface, in which users reformulate the original natural language questions in a simplified version of English, and a novel problem solver that can elaborate initially inadequate logical interpretations of a question by selecting relevant pieces of knowledge in the knowledge base. An evaluation of the work in 2006 showed that this approach is feasible and that complex, multisentence questions can be posed and answered, thus illustrating novel ways of dealing with the knowledge capture impedance between users and a formal knowledge base, while also revealing challenges that still remain.

In this paper, we study the problem of ordering subgoals under binding pattern restrictions for queries posed as nonrecursive Datalog programs. We prove that despite their limited expressive power, the problem is computationally hard — PSPACE-complete in the size of the nonrecursive Datalog program even for fairly restricted cases. As a practical solution to this problem, we develop an asymptotically optimal algorithm that runs in time linear in the size of the query plan. We also study extensions of our algorithm that efficiently solve other query planning problems under binding pattern restrictions. These problems include conjunctive queries with nested grouping constraints, distributed conjunctive queries, and first-order queries.

In order for a knowledge capture system to be effective, it needs to not only acquire general domain knowledge from experts, but also capture the specific problem-solving scenarios and questions which those experts are interested in solving using that knowledge. For some tasks, this latter aspect of knowledge capture is straightforward. In other cases, in particular for systems aimed at a wide variety of tasks, the question-posing aspect of knowledge capture can be a challenge in its own right. In this paper, we present the approach we have developed to address this challenge, based on the creation of a catalog of domain-independent question types and the extension of question template methods with graphical tools. Our goal was that domain experts could directly convey complex questions to a machine, in a form which it could then reason with. We evaluated the resulting system over several weeks, and in this paper we report some important lessons learned from this evaluation, revealing several interesting strengths and weaknesses of the approach.

Despite some successes, the lack of tools to allow subject matter experts to directly enter, query, and debug formal domain knowledge in a knowledge-base still remains a major obstacle to their deployment. Our goal is to create such tools, so that a trained knowledge engineer is no longer required to mediate the interaction. This paper presents our work on the knowledge entry part of this overall knowledge capture task, which is based on several claims: that users can construct representations by connecting pre-fabricated, representational components, rather than writing low-level axioms; that these components can be presented to users as graphs; and the user can then perform composition through graph manipulation operations. To operationalize this, we have developed a novel technique of graphical dialog using examples of the component concepts, followed by an automated process for generalizing the user's graphically-entered assertions into axioms. We present these claims, our approach, the system (called SHAKEN) that we are developing, and an evaluation of our progress based on having users encode knowledge using the system. Keywords Graphical knowledge entry, knowledge acquisition, components, composition, knowledge-based systems.

It was shown by Yannakakis that a locking policy is not safe if and only if there exists a canonical non-serializable schedule of transactions running according to the rules of the policy in which all the transactions except one are executed serially [Yan82]. In the present paper, we study the generalization of this result to a dynamic database, that is, a database that may undergo insertions and deletions of entities. We illustrate the utility of this generalization by applying it to obtain correctness proofs of three locking policies that handle dynamic databases. Keywords: Concurrency Control, Correctness Issues Safe Locking Policies for Dynamic Databases 1 1 Introduction A locking policy is called safe if any concurrent execution of a set of transactions while locked according to that policy is guaranteed to be correct. Yannakakis showed that a locking policy is not safe if and only if there exists a canonical non-serializable schedule in which all transactions except one ...

Large knowledge bases that are intended for applications such as CAD, corporate repositories or process control will have to be shared by multiple users. For these systems to scale up, to give acceptable performance and to exhibit consistent behavior, it is mandatory to synchronize user transactions using a concurrency control algorithm. In this paper, we examine a novel concurrency control policy called Dynamic Directed Graph (or DDG) policy that effectively exploits the rich semantic structure of a knowledge base.