Developers worldwide have been left astonished after only tinkering with Open AI’s GPT-3 natural language processing model, labeling it a vanguard product in the natural language processing (NLP) space.
Last year, OpenAI first described the state-of-the-art GPT-3 model in a research paper, “Language Models are Few-Shot Learners” which is available on a preprint server. GPT-3 is the largest NLP transformer ever at 175B parameters that is ten times the previous record held by Microsoft Research’s Turing-NLG at 17B parameters. Most AI systems are designed for one use-case; however, the GPT-3 is versatile. Its transformative components can be molded into any framework like a general-purpose “text in, text out” interface. The GPT-3 currently has four models: Davinci, Curie, Babbage, and Ada, each with different strengths and weaknesses. On another note, it is worth mentioning that the GPT-3 is not artificial general intelligence.
In May of 2020, at the Build developers conference, Microsoft CEO Satya Nadella introduced GPT-3’s first-ever commercial application: Microsoft Power Px formulas is based on natural language input using a fine-tuned version of GPT-3. As of March of 2021, the GPT-3 model generates an average of 4.5B words per day in its over 300 applications which harness GPT-3’s capabilities.
Image credits: Open AI
Applications of GPT-3
GPT-3’s applications dwell in various industries ranging from productivity and education to creativity and games. GPT-3 advances most text-based operations that require some sort of understanding.
The GPT-3 model is being used to augment mental health therapy online. Koko is a platform that is based on the concept of crowdsourced cognitive therapy. Koko identifies Individuals at risk of self-harm and suicide using NLP and deep learning.
Another startup, Curai Health, expands healthcare access by offering chat-based primary care at a lower cost. They amalgamate AI and experts, an “AI with experts in the loop” approach. The experts verify the AI’s output and make sure it is appropriate.
Algolia redefines semantic search using the GPT-3 model. The service incorporates deeper contextual information understanding to enhance searching on websites, incorporating dynamic re-ranking, query suggestions. Moreover, Algolia offers analytics on search, allowing businesses to adapt accordingly.
ActiveChat Inc provides natural language understanding chatbots capable of live chat CRM, conversational intelligence, and general-purpose chatbot. ActiveChat allows users to build streamlined flows using a visual conversation builder, described as “LEGO for chatbots.” Lastly, all chat history is analyzed, giving insights into customer sentiment tracking, intent analytics, and conversion rate optimization. There are other similar options, for instance, Viable, Sapling, and MessageBird.
Conto AI uses natural language understanding to aid content creation. They employ NLP to generate social media content for a Facebook ad description, LinkedIn headline, Instagram caption, etc. The platform asks the user to input basic details such as brand or product name, audience, and description to generate content. There are many similar platforms, such as Copysmit, Copy.AI, and Replier.ai.
General Business Productivity
OthersideAI’s text system currently uses GPT-3 to generate essays, blogs, and email responses. OthersideAI’s system learns a user’s writing style and asks for context like an introductory paragraph and points which will be included in the text. The AI can infer the temporal tense from brief prompts, for instance, turning ‘see you tomorrow’ into ‘I will see you tomorrow.’
As the NLP market grows from three billion U.S. dollars in 2017 to over a predicted 43 billion in 2025 . Usage of models such as the GPT-3 will only increase, propelling the growth of the NLP market and businesses in general. It is pivotal that businesses remain agnostic with the technologies they use and always look out for newer and more advanced inventions.
- Liu, S. (2021, October 19). Global Natural Language Processing Market 2017-2025. Statista. Retrieved December 13, 2021, from https://www.statista.com/statistics/607891/worldwide-natural-language-processing-market-revenues/.
About the author: Shloak Rathod is an open-source developer and researcher interested in AI, web technologies, data science, and economics. He relishes writing about fast-paced upcoming technologies with disruptive capabilities. He also loves a chat, don’t hesitate to reach him at firstname.lastname@example.org.