Intel Presents NLP Architect with Several Features to Augment NLP Research
A team of researchers and developers at Intel AI Lab has launched “NLP Architect”, an open-source Python library for NLP. The new library enables data scientists, researchers and developers to explore advanced deep learning techniques in the field of natural language processing and natural language understanding.
According to Intel’s recent blog post, the current version of NLP Architect offers a set of features for both research and practical applications:
NLP core models that allow robust extraction of linguistic features for NLP workflow: for example, dependency parser (BIST) and NP chunker
NLU modules that provide best in class performance: for example, intent extraction (IE), name entity recognition (NER)
Modules that address semantic understanding: for example, colocations, most common word sense, NP embedding representation (e.g. NP2V)
Components instrumental for conversational AI: for example, ChatBot applications, including dialog system, sequence chunking, and IE
End-to-end DL applications using new topologies: for example, Q&A, machine reading comprehension
“We look at the above features as a set of building blocks that are needed for implementing NLP use cases based on our pragmatic research experience”, states Yinyin Liu, Head of Data Science and AI Products Group at Intel. Liu also says that “the team has put together a set of deep learning-driven NLP models, not specific to a particular domain or application, but are ready to look at potential use cases, and use some of these building blocks in the library”.
With NLP Architect, researchers and developers could infuse smart capabilities into conversational agents and chatbots, such as name entity recognition, intent extraction, or semantic parsing. This in turn allows the system to automatically identify a person’s upcoming actions, based on their words.
Since its launch in December 2017, Intel AI Lab has open-sourced libraries to support researchers in deploying reinforcement learning and neural networks. Some of the leading examples include neural network distiller library (used to get rid of neural connections irrelevant for a task) and Coach, the reinforcement learning library (used to embed an agent in training environments like robotics or self-driving car simulators).
NLP Architect was launched by Intel during its inaugural AI DevCon event, held last week at San Francisco. The conference hosted several top minds in AI, machine learning, deep learning and data science.