Careers in AI — AI in Research
Nowadays, AI research is done not only by traditional research institutions but also organisations across numerous different sectors. This article is part of the ‘Careers in AI’ series, and in this article we explore careers in traditional Artificial Intelligence research.
Authors: Jenny Shim
What is AI research about?
AI research impacts technology used by people in their daily lives, from conducting fundamental research to influencing project development. The industry makes discoveries that impact everyone and shares their research and tools to fuel progress in the field. Some researchers publish regularly in academic journals, release projects as open source, and apply research to their products. The field lies between academia and industry and provides pathways for development and introduction of new technologies to industrial problems.
Why consider AI research?
If you are passionate about making discoveries that impact everyone and fuelling progress in the field of AI excites you, then AI research may be a good fit for you. Being an AI researcher allows you to gain access to the newest technologies and findings which means that every day you will be learning new things and achieving breakthroughs. An inquisitive, creative person will find the work very rewarding and intriguing as the work involves discussing new ideas, carrying out research and solve novel, untouched problems.
As innovation in AI grows, there is an opportunity — and responsibility — to ensure that artificially intelligent systems are built to contribute to the public good and to a well-functioning economy, with fairness, reliability, security, and appropriate transparency and privacy at its core (Alan Turing institute).
Fields of Artificial Intelligence
There are countless areas on can go into in the field of research. Below you may find some of the most popular research areas, which include algorithms, data mining, robotics, NLP and privacy.
Algorithms and Theory
The industry presents many exciting algorithmic and optimisation challenges across different product areas including Search, Ads, Social, and Infrastructure. These include optimising internal systems such as scheduling the machines that power the numerous computations done each day, as well as optimisations that affect core products and users, from online allocation of ads to page-views to automatic management of ad campaigns, and from clustering large-scale graphs to finding best paths in transportation networks. Other than employing new algorithmic ideas to impact millions of users, researchers contribute to the state-of-the-art research in these areas by publishing in top conferences and journals.
Data mining and modelling
The proliferation of machine learning means that learned classifiers lie at the core of many products across the industry. However, questions in practice are rarely so clean as to just to use an out-of-the-box algorithm. A big challenge is in developing metrics, designing experimental methodologies, and modeling the space to create parsimonious representations that capture the fundamentals of the problem. These problems cut across products and services, from designing experiments for testing new auction algorithms to developing automated metrics to measure the quality of a road map. Data mining lies at the heart of many of these questions, and its research is at the forefront of the field. Whether it is finding more efficient algorithms for working with massive data sets, developing privacy-preserving methods for classification, or designing new machine learning approaches, the industry continues to push the boundary of what is possible.
Robotics research has significantly advanced ourn knowledge base. The future of robotics research and its impact largely depends on broader collaboration between researchers in various fields (e.g. engineering, science, arts, business, law etc.) and end-users in various market domains (e.g. the manufacturing industry, healthcare, agriculture, consumer, civil, commercial, logistics and transport). Such collaboration has the potential to generate huge impact by developing solutions to end-user driven challenges, which will also advance the state of the art of robotics research. Substantial progress in robotics research is being made very quickly. On the other end of the spectrum, the pace of applying intelligent and autonomous robots in real world applications is slow. Not many intelligent/autonomous robots have been practically deployed or available in the market. In particular, there is a huge gap between Technology Readiness Levels (TRL) 6 and 7, at the transition between research and technology development on one hand and product and solution development on the other. Part of the reasons for this could be that robotics researchers may not fully appreciate the challenges related to applying research in real world applications, and the collaboration between academia and end-users is not appropriate enough.
Having a machine learning agent interact with its environment requires true unsupervised learning, skill acquisition, active learning, exploration and reinforcement, all ingredients of human learning that are still not well understood or exploited through the supervised approaches that dominate deep learning today.
The main goal is to improve robotics via machine learning and improve machine learning via robotics. The industry fosters close collaborations between machine learning researchers and roboticists to enable learning at scale on real and simulated robotic systems.
Natural language processing
Natural language processing (NLP) is a subfield of linguistics, computer science, and artificial intelligence concerned with the interactions between computers and human language, in particular how to program computers to process and analyse large amounts of natural language data. The result is a computer capable of “understanding” the contents of documents, including the contextual nuances of the language within them. The technology can then accurately extract information and insights contained in the documents as well as categorize and organize the documents themselves. Challenges in natural language processing frequently involve speech recognition, natural language understanding, and natural-language generation.
Natural Language Processing (NLP) research at Google focuses on algorithms that apply at scale, across languages, and across domains. Their systems are used in numerous ways across Google, impacting user experience in search, mobile, apps, ads, translate and more. Their work spans the range of traditional NLP tasks, with general-purpose syntax and semantic algorithms underpinning more specialised systems. They are particularly interested in algorithms that scale well and can be run efficiently in a highly distributed environment. Recent work has focused on incorporating multiple sources of knowledge and information to aid with analysis of text, as well as applying frame semantics at the noun phrase, sentence, and document level.
Active research is being done exploring virtually all aspects of machine learning, including deep learning and more classical algorithms. Exploring theory as well as application, much of the work on language, speech, translation, visual processing, ranking and prediction relies on Machine Intelligence. In all of those tasks and many others, large volumes of direct or indirect evidence of relationships of interest are gathered, applying learning algorithms to understand and generalise.
Machine Intelligence raises deep scientific and engineering challenges, allowing researchers to contribute to the broader academic research community through technical talks and publications in major conferences and journals. Contrary to much of current theory and practice, the statistics of the data observed shifts rapidly, the features of interest change as well, and the volume of data often requires enormous computation capacity. When learning systems are placed at the core of interactive services in a fast changing and sometimes adversarial environment, combinations of techniques including deep learning and statistical models need to be combined with ideas from control and game theory.
Security, privacy and abuse prevention
The Internet and the World Wide Web have brought many changes that provide huge benefits, in particular by giving people easy access to information that was previously unavailable, or simply hard to find. Unfortunately, these changes have raised many new challenges in the security of computer systems and the protection of information against unauthorised access and abusive usage. For example, at Google, their primary focus is the user, and their safety. People work on nearly every aspect of security, privacy, and anti-abuse including access control and information security, networking, operating systems, language design, cryptography, fraud detection and prevention, spam and abuse detection, denial of service, anonymity, privacy-preserving systems, disclosure controls, as well as user interfaces and other human-centred aspects of security and privacy. Their security and privacy efforts cover a broad range of systems including mobile, cloud, distributed, sensors and embedded systems, and large-scale machine learning.
Personal financial management has also been a particularly promising space. Some platforms are built to help users manage their personal finances. These tools are highly customisable, and offer the ability to make recommendations, create spending or savings goals, and plan/execute short- and long-term tasks, from paying bills to preparing tax filings. Mint and Wallet are examples of companies dedicated to offer personal financial management tools.
Types of companies
Mergers and acquisitions happen very actively in this industry, mostly in a form of bigger companies like Google acquiring startups. Some of the well known organisations are listed below.
● Corporations — Google, Microsoft, Amazon, Apple, IMB, Intel, Qualcomm
● Research organisations — The Alan-Turing Institute, Open AI, Artificial Intelligence Centre, Association for the Advancement of AI
● University labs
Types of Jobs
Whether developing experiments, prototyping implementations, or designing new architectures, research scientists work on a real-world problem in computer science. Research scientists work across data mining, natural language processing, hardware and software performance analysis, improving compilation techniques for mobile platforms, core search and much more.
The research-focused software engineers set up large-scale tests and deploy promising ideas quickly and broadly. From creating experiments and prototyping implementations to designing new architectures, research engineers work on machine learning, data mining, hardware and software performance analysis, improving compilers for mobile platforms and much more.
A Data Scientist will evaluate and improve different products. They will collaborate with a multi-disciplinary team of engineers and analysts on a wide range of problems. This position will bring scientific rigour and statistical methods to the challenges of product creation, development and improvement with an appreciation for the behaviours of the end user. Data scientists not only revolutionise search, they routinely work on massive scalability and storage solutions, large-scale applications and entirely new platforms for developers around the world.
Professorship (Computer Science)
An academic body will be a member of a University community; part of a lively and intellectually stimulating research community with access to the excellent research facilities which the university offers. They must hold a doctoral degree in Computer Science (or cognate discipline), have the ability to teach across a range of Computer Science subjects, and also have a proven research record of high quality at international level in the area of Artificial Intelligence, and experience of research collaborations at both national and international level. They will secure research funding, engage in the management of research projects, and disseminate research of the highest international standard through journals, conferences and seminars, and teach on the department’s undergraduate and/or postgraduate courses.
Postdoctoral Fellow (Healthcare)
This position provides the opportunity to work with leading drug discovery and AI scientists, with access to unique sets of proprietary data to assist in the identification of new drug molecules. They will focus on developing deep learning AI models to enable the prediction of compound activity against a specific class of targets. This will entail compiling structure-activity datasets from internal and external data sources and utilizing cutting-edge deep learning methodology to construct models to predict the activity profile of existing and novel molecules across multiple, related targets. They will develop novel approaches and methodology to enable the inclusion of protein structural information into deep learning models and will be expected to publish and present their findings at internal and external meetings. During this postdoc, they will be immersed within a vibrant and productive chemistry group in a successful oncology drug discovery department. The project will be supervised by both researchers at a company and research scientists of a university, allowing them to benefit from leading academic and industrial research environments.
ML engineer (Healthcare)
This is a hands-on position where you will be empowered to be creative, ambitious and bold, to solve novel R&D problems and have the potential to directly impact the lives of patients living with disease. A machine Learning Engineer will Identify opportunities to apply the latest advancements in NLP and Machine Learning to build, test, and validate models in named entity recognition, relation extraction, and entity linking. They will also promote and use standard biomedical terminologies and ontologies (e.g. UMLS) and automatically extract knowledge from the scientific literature to be combined with internal and external datasets as a biomedical knowledge graph. There will be opportunities to connect and collaborate with subject matter experts in biology, chemistry, and medicine.Example of interview questions
Examples of Non-Technical questions:
Research background and knowledge
Major contribution within the field
Situational and competency questions
Examples of Technical questions:
Computer science fundamentals and programming e.g. binary trees
Probability and statistics
Data modelling and evaluation e.g. classification of machine learning
Application of machine learning algorithms and libraries
Software engineering and system design
Links & References