Anil, thank you for taking the time to meet with me today – it would be great if you could start by introducing yourself?
Thanks Harry. I’m a data professional with a PhD in Natural Language Processing and I’m passionate about creating real-world impact, through solving business problems using cutting-edge AI solutions. My main areas of interest include NLP, Machine Learning, Deep Learning and AI. I’m currently the Head of Data Science at Logically, leading the efforts in building AI-based solutions to combat misinformation and fake news.
Amazing! What were you doing before Logically?
Before joining Logically, I was working as an NLP Data Scientist at a CX analytics startup called SentiSum, where I was the first Data Scientist in the business. My job was to enhance the capabilities of their NLP engine that would process, analyse and understand customer feedback data mainly in the form of online reviews. I worked closely with the back-end, front-end engineers and product owners to understand the business problems and in turn develop NLP solutions.
Why did you leave SentiSum to join Logically?
At SentiSum I had the opportunity to build NLP solutions to understand textual customer feedback, and whilst I enjoyed my work at SentiSum, I was looking for a new job that could give me an opportunity to make a greater impact. Logically’s mission and vision to tackle online misinformation and their AI-based efforts to combat it caught my attention. I was highly motivated after hearing from Lyric, the CEO, about how Logically is developing cutting-edge, AI based solutions to combat information overload, misinformation, fake news and other types of harmful content.
How would you describe what Logically do?
Logically is a technology company specialising in the use of artificial intelligence (AI) to improve the consumption and interactions with information online. We develop both B2C and B2B products that use AI to analyse the vast quantities of information on the internet and assess its credibility and veracity.
Fundamental to our mission is to improve civic discourse. Our B2C products are developed to provide users with credible analysis of the news content that they are reading online. Through a combination of cutting-edge AI-driven insights and human-centric fact-checking operations, we empower our users to take a more holistic approach to their news consumption habits. Our B2B products empower public and private organisations with intelligence and insights to combat misinformation and improve decision making.
How are Logically currently using AI?
AI forms the core of Logically’s solutions to combat misinformation and fake news detection. All our products and tools that offer a unique experience to our users/clients for news consumption, discussion and verification are built leveraging the insights from our machine learning models. Also, we are working on making our models more interpretable and explainable, which is the latest trend in the industry. We envision to make our solutions augmented (human + machine) intelligence-based to best avoid bias and to gain continuous improvements in accuracy.
How do you and your team fit into all of this?
Logically’s Data Science team along with the Engineering team builds the technology behind our products/tools. A variety of our product features (e.g. summary of a news story) are entirely AI-based. Our Data Scientists work on business problems and develop effective solutions using cutting-edge AI and machine learning. My role as the Head of Data Science is to ensure my team’s data science contributions align with the company’s mission and vision.
How does the use of AI benefit your platform?
Our platform immensely benefits from AI. From the moment data is ingested into our databases, our machine learning models kick in to understand the underlying patterns for news insights generation. Currently we have up to 25 different AI models that make our platform smart and the number is increasing constantly.
Could you tell us a little more about a few of the models you’re currently using?
Sure. We have built models that give state-of-the-art performance for tasks ranging from named-entity recognition, sentiment analysis to story summary generation to hate-speech and profanity detection. We constantly monitor the performance of our models in production and invest in new approaches that are setting new industry standards to improve the quality of our models.
What are Logically’s future plans and how does AI fit into it?
Logically’s future is to enter B2B and B2G markets with innovative solutions to address business and government level problems with fake news. Whilst our current B2C news app is being used by a growing user base, we are planning to leverage the success of our solutions to build several other digital tools to combat misinformation that can integrate easily with third-party tools and frameworks. Here again, AI will be heavily leveraged for analysis and insights of enterprise and government data about harmful and problematic content.
What are you most excited for?
I’m very excited about advancing the AI-efforts within the company and also equally looking forward to our upcoming projects in the areas of explainable-AI and augmented (AI + Human) intelligence. Further, I’m highly excited about the success of our recent projects with election-related misinformation monitoring in India, the UK and the upcoming launch of our B2C app in the US for the presidential elections 2020. Furthermore, the growing potential of Logically to become the disruptive leader in combating fake news and creating a greater impact on our society gives me immense satisfaction.
Lastly, what is one book would you recommend?
“The Magic of Thinking Big” by David J. Schwartz.