Microsoft’s Research and AI

​The world of computing is such an interesting and important field because of its never-ending nature and possibilities. Starting with the first computer built in 1943, computing has advanced far beyond what anyone at the time would have imagined. This is a result of decades of technological research and companies capitalizing on said research. One of these institutions is Microsoft. Known for being a leading vendor in computer software, hardware for pc, mobile, and gaming systems, they have become one of the most influential and important technology companies in the world. One field that Microsoft specializes in that most people don’t think of first, is AI, mostly in part due to all the other accomplishments. Microsoft has led in the AI patent race, as a result of their research subsidiary known as Microsoft Research. AI, which stands for “Artificial Intelligence”, is a branch of computer science that is concerned with smart computers that can perform tasks that require human intelligence. Microsoft Research has contributed and made significant advances in AI, implemented into its products: Bing, Cortana, Linkedin, and many more.

​​When thinking of search engines, Google is probably the one that pops up for most, but Bing has been a silent competitor for years while competing against many other search engines. So, what sets Bing apart? For Bing, they had to find a way to have better search results, satisfying the needs of the hundreds of millions of users using their search queries. Over the past few years, Bing research has worked with Microsoft Research to develop large network models such as MT-DNN, Unicode, UniLM, to maximize user experience. The result of these AI at scale is a new intelligent answer, expansion of intelligent answers globally, and improving search relevance. Before, when users wanted to confirm an answer to their question, the search result would have looked through many documents to find the best answer, but this new intelligent answer has advanced far enough where it can answer questions with a yes/no statement, being more efficient to users.

Before:

After:

AI has also allowed for advancement in non-English speaking markets. Using multi-task deep learning, Bing has been able to translate and generate intelligent answers, using the same universal model that is operated in English-speaking markets, allowing people that do not speak English to get the same experience.

Query in Italian:

The English translation of the Italian query:

Lastly, AI has helped in understanding search intent, improving search relevance. The new model can understand complex and ambiguous concepts, such as assuming if a user mistyped or misremembered something, and fine-tuning search results. All of these are examples of the capabilities and experiences of AI, and with the help of Microsoft, Bing was able to improve fulfilling the needs of its users. In terms of success, Bing has been able to gather more total numbers of visits since the introduction of the project. Seeing a slow decrease of visits from May 2019 to early 2020, Bing has thrived and has started gaining more visits since the introduction of Microsoft Research’s AI at scale.

​ ​ ​​Linkedin, a company under Microsoft, has helped many find jobs in their respective fields with the help of AI. Using all users’ work history, skill sets, or connections, Linkedin’s AI has improved user experience for all, whether it be recommending people to connect to or recommending articles and updates on the news feed.

​ ​ ​​If you are a recruiter, it recommends candidates that should be sourced to fill up roles. If you are looking for a job, it recommends jobs you should apply for, as well as recommending courses to take. For sales, it recommends people that can close out a deal for you. How successful has AI been for LinkedIn? Since the improvement of LinkedIn AI tools, there was a 30% increase in job applications, email responses have increased by 45%, improving article recommendation click-rate by 10–20%, and having engagement increased by 50% year after year. LinkedIn has quickly become the most advantageous social networking tool used by professionals and will only continue to grow with the help of artificial intelligence.

​ ​ ​​Lastly, Cortana. Cortana is Microsoft’s AI-driven virtual assistant, whose goals are to aid Windows 10 users in pledge requirements, accomplish chores, and use applicable data in a personal context. Microsoft Cortana uses Bing’s search engine to execute simple tasks and is accessible in different languages. Other features include being able to be used with third-party apps on Windows 10 or openly over the service. Before Cortana, Apple users had Siri, which allowed for personal assistants on personal computers and smartphones. Cortana was a response to Apple and was created for the same use. Having such software anywhere you go makes everything more efficient, allowing users to set reminders, make phone calls, search the web, without having to directly access the web. However, Cortana has not been as successful as its competitors. In early 2021, Microsoft ended support for Cortana on Android, IOS, and a couple of other small platforms. To no one’s surprise, Cortana could not compete with Google Assistant on Android phones or Siri for IOS users, or even Amazon Alexa. While unsuccessful, Cortana has laid the groundwork for the future of AI. Microsoft’s Azure Bot Services and Bot Framework are among the most popular for chatbot development, and Microsoft can very well be the leader of voice and conversational AI.

​ ​ ​​In conclusion, Microsoft Research has been at the center of ubicomp research and systems. AI is just one of the features that Microsoft is the leader of, and it has greatly benefited society. By implementing an incredibly advanced AI to Bing’s search query, users can have more beneficial answers. In Linkedin, users, whether it be recruiters or career seekers, can connect to people more relevant to them. Cortana has much to offer, and the world of AI virtual assistants has just begun and has the potential to reshape society.

​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​

​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​​ ​ ​​ Works Cited

Babi, Jona. “LinkedIn — Ai Helps Companies and Professionals Find Each Other Faster.” Digital Innovation and Transformation, digital.hbs.edu/platform-digit/submission/linkedin-ai-helps-companies-and-professionals-find-each-other-faster/.

Columbus, Louis. “Microsoft Leads the AI Patent Race Going into 2019.” Forbes, Forbes Magazine, 7 Jan. 2019, www.forbes.com/sites/louiscolumbus/2019/01/06/microsoft-leads-the-ai-patent-race-going-into-2019/?sh=296a287a44de.

“Emerging Technology, Computer, and Software Research.” Microsoft Research, 6 Oct. 2021, www.microsoft.com/en-us/research/.

Gao, Richard, and Richard Gao (1932 Articles Published) . “Microsoft Is Basically Giving up on Cortana, Android Support To Die Next Year.” AndroidPolice, 1 Aug. 2020, www.androidpolice.com/2020/08/01/microsoft-is-basically-giving-up-on-cortana-android-support-to-die-next-year/.

Gregsays: et al. “Microsoft Bing Usage and Revenue Stats (New Data).” Backlinko, 5 Mar. 2021, backlinko.com/bing-users.

“Microsoft Announces New Supercomputer, Lays out Vision for Future AI Work.” The AI Blog, 20 May 2020, blogs.microsoft.com/ai/openai-azure-supercomputer/.

Pratap, Abhijeet. “Microsoft Research and Development Expenses.” Notesmatic, 26 Feb. 2021, notesmatic.com/microsoft-research-and-development-expenses/.

theDevMasters. “Ai Tools (Microsoft Cortona).” AI Tools (Microsoft Cortona), www.thedevmasters.com/ai-tools/microsoft-cortana.