As interested-in-language-processing guy I visited presentation of Goldfire product.
Goldfire is The Platform for Optimal Product Decisions, as the page says. And I’m absolutely sure that it’s fulled with nlp tasks solutions (patent links [1] [2] [3], etc.)
The presentation can be divided into three parts: introduction, demonstration, q-a session.
During the introduction it was told a little about history of product and main concepts of any natural language processing product
Goldfire screenshot from the site
Parameters (main parts of the nlp system)
- Knowledge base
- Data accumulation & learning
- Derivation (solution generation)
- Natural communication
The development includes tagged corpora, precision-recall curve for quality valuation, lexical-grammar analysis + semantic + concept.
The internal Question-answering system works by sorting question types, e.g. Why … cause, How to … cause.
The internal system innovation is semantic lenses, special semantic filters.
Some features are root cause analysis, causes and effects search, trimming, modeling search tools, cross-lang search+translation, sentiment analysis.
Plans for future are:
- knowledge inference
- decision making
- long-short term memory
- more applications, more langs
- permanent quality improvement
btw, chinese was added last year
Demonstrations can be found on Youtube channel.
p.s. Goldfire is expensive software and most suitable for big companies.
upd. short list of customers: Boeing, Schneider Electric, NASA, Johnson & Johnson, Bosch, Chrysler, Ford, Hyundai, Nestle, Royal Canin, LG, Samsung, Sony, Toshiba, Georgia Institute of Technology, University of Connecticut, University of Missouri
