|
|
Nervana ~ Discovery Zone
 Nervana Discovery Solution Comparison Samples 
Ontology
An ontology is a computer-readable taxonomy of concepts (diseases, proteins, clinical tests) linked by relationships (is a kind of, is a part of, has function) that encodes human knowledge for a particular domain. Ontologies can represent immensely complex knowledge in a certain domain (such as anatomy, or manufacturing).
|
|
Taxonomy
A taxonomy is a hierarchical classification for a subject or domain (such as Life Sciences or Finance) organized into categories and subcategories. A taxonomy provides a way to separate entities into buckets or relatively broad topic levels.
Typically, taxonomy structures are developed by taxonomy specialists or respected institutions serving as domain specialists. Additionally, companies often develop in-house taxonomies to reflect categories relevant to their business, such as their products, major projects, industries served, and so on.
|
|
|
Semantics
Semantics is the study of meaning at the sentence level; the context of what words mean when used in a sentence.
For example, "Dr. X is the chair of the ABC committee". In this sentence, the semantic context of the word "chair" is obviously not referring to a piece of furniture. Instead it refers to his role as chairman. There may also be other ways to describe his role (such as leader, decision maker, chief advisor.)
To allow Nervana to recognize multiple descriptions for a given word or phrase, the system uses ontologies to gain a semantic understanding of the meanings of words and phrases as they relate to specific domains of knowledge.
|
|
|
Semantic network
A semantic network structures concepts and terms as a network or a web, rather than as a hierarchy. The Nervana semantic network contains a database of documents, entities, experts, and events. This semantic network is then able to be dynamically queried.
|
|
|
Semantic Wildcard
The Nervana System contains Semantic Wildcards; a powerful way to search for concepts and obtain results that are relevant to a precise concept. The semantic search returns information that you may not know, relevant to a concept. This approach brings up information that otherwise would have been missed using a keyword search. Nervana’s Semantic Wildcard is a “*:” (star colon) used at the front of a term or concept.
For example, entering *:cancer returns articles that contain the phrase cancer, along with articles containing references to lymphoma, leukemia, sarcoma, carcinoma; any type of cancer even if the actual word cancer is not mentioned.
|
|
|
Semantic ranking
Semantic ranking is a Nervana-patented process where documents are analyzed and compared to each other. Complex mathematical mappings build a multi-dimensional representation of each document on the fly. Each document summary extracted by the Nervana System is then provided a semantic weight relative to its semantic profile. Semantic weighting is the percentage given to a document in comparison to all other documents in the index. The percentage is based on the profiling of the document. Additional ranking techniques include multi-dimensional ranking, context sensitivity, semantic density and co-location, timeliness, and adaptive ranking methods.
|
|
|
Knowledge discovery
Nervana servers semantically analyze and index the information, creating a database to contain the results. This database contains a semantic network of documents, entities, experts, and events. This semantic network is then able to be explored dynamically, providing a more natural and meaningful interaction between the user and the materials they are searching.
As a result, knowledge discovery is condensed to mere seconds and results are more meaningful. These efficiencies can make a significant difference to a company’s business and revenue.
|
|
|
Natural Language Query
Nervana allows you to ask questions “naturally” and within the context of their meaning, crossing multiple domains and information repositories. For example, a researcher can ask: “Find competitive intelligence on drugs that target obesity or other metabolic disorders in teenagers and children.” You would then search the Nervana System using the following combination: (*:obesity OR “*:metabolic syndrome”) (*:teenagers OR *:children) *:drugs. Nervana doesn’t just extract keywords from the query, rather it sifts through all of the possible meanings and contexts for this request and returns the most relevant, timely results from the system.
|
|
|
Dossier
A Dossier displays an overview of results grouped by different Knowledge Types, such as Best Bets, Breaking News, and All Bets—that are relevant to your request.
|
|
|
Knowledge Community (KC)
Information sources that have been crawled and indexed are then grouped into a Knowledge Community focused on a specific knowledge domain such as Biotech or Finance, or on a major content resource, such as MEDLINE®. When you create a profile, KCs are added so that Nervana knows which sources to search.
|
|
|
Profile
Profiles represent your various search scenarios. They also enable you to take advantage of the power of federation in your search. You can create multiple profiles, each pointing to different combinations of KCs. For example, a Business Discovery profile includes the Life Sciences News and General News KCs. A Drug Discovery profile includes the Life Sciences News, General News, and Life Sciences Patents KCs.
|
|
|
Federation
The Nervana System allows you to search across multiple sources. Nervana federates across sources in real time, as well as across categories of information. Federated search saves time by eliminating the need to search multiple sources individually. Nervana also federates across ontologies, while still keeping them as individual entities.
|
|
|
KIS (Knowledge Integration Server)
The Nervana KIS crawls and semantically analyzes data sources (such as web sites, file shares, and databases) to create a semantic index of all of the content. As information in any of these data sources is updated or changed, the KIS also automatically detects changes and updates the index. The KIS communicates with the KDS for document analysis and builds a semantic network. The semantic network contains links for each document summary.
|
|
|
Knowledge Domain Server (KDS)
The Nervana KDS hosts one or more ontologies that describe the semantic vocabulary of various knowledge domains, as well as other linguistic components and algorithms. The KDS sends semantic information to the KIS describing what a document, text, or list of concepts means after having categorized the information according to one or more ontologies.
|
|
|
Alerts
Receive notification by e-mail of new articles or items of interest on your topic or research interest.
|
|
|
Metadata
"Data about data"; information that describes the object you are currently working with.
Examples of metadata are names of fields, the kind of values that are allowed, the range of the values, explanations of what the fields mean, file attributes (name, size, data type), data about records (length, fields, columns), and data about data (where it is located, how it is associated, who owns it).
|
|
|
Connectors
Connectors point to internal document stores, email, content repositories, databases, and so on. Connectors allow Nervana to communicate with various industry standard content repositories. Nervana connectors integrate with major data sources such as Lotus Notes, Outlook, Oracle 9i and later, Microsoft SQL Server, and Documentum Content Management Server. Nervana never takes the information out to store it somewhere else. The system creates pointers to the original source.
|
|
|
Live Mode
A way to set up automatic queries, alerts, watch lists, and real-time results streaming to a customized view. View options include a ticker tape, sidebar, or screensaver.
|
|
|
“Drag-and-Drop” queries
Use a document, piece of text, or an entity to create a query, where the system analyzes the sample and then finds semantically similar materials. The Nervana System analyzes the document (and its various points of view) retrieving everything out of the semantic network that is related to the original document. For example, you could use a competitor’s press release to monitor activity in the press around their latest announcements.
|
|
|
|