NLP consultant

Examples of what an NLP consultant can do

Examples of applications of natural language processing include:

• Natural language understanding
• Text generation including LLMs such as LLaMa or GPT-4: we’ve all used GPT in the browser. But NLP consultants can train custom LLMs using a domain-specific or language-specific corpus, use prompt engineering to get an LLM to perform a domain specific task such as answering questions about English insolvency law, or leverage the data returned by an LLM such as token scores to calculate perplexity, identify if a text has been written by an LLM, or detect plagiarism
• Identify the author of an unknown text (forensic stylometry)
• Natural language dialogue systems - for example, develop a multilingual chatbot specifically for a company and domain, with detailed knowledge of a particular product range. We can leverage state of the art large language models (LLMs) such as GPT-3, or even train our own LLM.
• Text analysis - which molecules were used in a biosciences paper? How many subjects in a clinical trial? What percentage of doctors felt there was an interaction between Drug A and Drug B?
• Multilingual models capable of handling survey data from different countries
• Predictive models based on text input
• Topic analysis – clustering, finding similar documents
• Domain specific named entity recognition such as drug named entity recognition and linking, company named entity recognition, or country named entity recognition and linking.
• Document classification: is an incoming email a request to change address, deregister a company, or a request for a VAT invoice? Is a customer complaint high risk and needs to be dealt with immediately, or lower priority (triage)?
• Document-based recommender systems
• Unstructured data analysis
• Document anonymisation: can we remove all GDPR- or HIPAA-sensitive data from a PDF? Can we train a machine learning model from sensitive data?

What would an NLP consultant do in the healthcare domain?

AI and natural language processing are being increasingly adopted across the healthcare sector. This technology is sometimes called healthtech or MedTech. NLP is being used to compare and detect changes in clinical reports, extract clinical concepts such as MeSH terms from electronic medical records, and develop human-to-machine natural language dialogue systems to improve the healthcare experience.

We have worked on a number of projects in healthcare, including:

• a model to predict the complexity of clinical trials from the trial protocol for Boehringer Ingelheim.
• a model to predict the risk of a clinical trial from the trial protocol for the Gates Foundation. Published in Wood TA and McNair D. Clinical Trial Risk Tool: software application using natural language processing to identify the risk of trial uninformativeness. Gates Open Res 2023, 7:56 doi: 10.12688/gatesopenres.14416.1
• a desktop application to analyse researchers' outputs, fields, collaborations and affiliations using PubMed search result exports.
• a model to identify researchers who have used open sourced molecules in their published research without attribution, also for Boehringer Ingelheim.

​What technologies does an NLP consultant use?

We do a lot of natural language processing with Python. We have worked on a variety of NLP models, including:

• Transformers, including custom trained transformers using libraries such as HuggingFace, Keras or spaCy
• Large Language Models (LLMs) such as Open AI’s GPT-3.5, Google BARD, or similar
• Bag of words, tf*idf, Jaccard similarity, cosine similarity
• NLP pipelines, lemmatisation, parsers, chunkers
• Deep neural networks
  • convolutional neural networks (text as well as images)
  • RNN, LSTM
  • Seq2seq, word2vec, doc2vec
  • see a live demo of a CNN for author identification
• Clustering: Latent Dirichlet Allocation
  • This is useful for extracting topics from a set of unstructured documents, for example legal documents, survey responses, factory error reports, etc.
• Search engines and search term recommenders
• Google Cloud, AWS, Microsoft Azure, Azure Cognitive Services, etc

​Natural Language Processing in Python and R

We work with the following programming languages and frameworks:

• Python
• TensorFlow
• spaCy
• Keras
• Natural Language Toolkit (NLTK)
• R

​Examples of past NLP consulting engagements

NLP consulting projects we have worked on for major household names include

• a spoken dialogue system to control a smart home
• an unsupervised text analysis program to analyse text descriptions of manufacturing defects (Boehringer Ingelheim)
• a model to classify jobseekers' CVs into industries and salary bands (CV-Library).
• analysis of survey responses and deployment of public NLP dashboards allowing members of the public to analyse and explore text datasets (White Ribbon Alliance)