Interview with Matthew Pallone, CTO at NeuroSoph

At the upcoming AI for Government Summit in Toronto, NeuroSoph will be joining as an exhibitor to showcase their work. NeuroSoph is currently developing an Intelligent Forms Processing System called Specto. Specto augments the processing of the needed paper-based and digital forms. The system extracts and classifies data from incoming documents and applies that output to downstream systems and workflows. Using AI, Machine Learning, and OCR technologies Specto integrates seamlessly into any workflow. We spoke with Matthew Pallone, CTO at NeuroSoph to learn more about what they were doing.

Give us an overview of your role as CTO at NeuroSoph.

NeuroSoph is an early-stage startup company, so everyone on the team is required to wear many hats! I work closely with our CEO and the other founders to identify current industry needs, design solutions, and lead our research and development efforts. I am primarily responsible for: 1) Keeping abreast of the latest developments in AI – especially with respect to advances in image processing, NLP, and NLU – and brainstorming ways in which those innovations could enhance our own business solutions. 2) Maintaining a research, design, and development methodology within the company, so that our projects remain focused, well-organized, and successful. 3) Working with our researchers and software developers to help them solve challenging engineering problems and keep them appraised of relevant resources, publications, and innovations. 4) Speaking with clients and potential clients to understand their needs and to outline efficient, achievable solutions.

How did you begin your work in AI?

In graduate school, I studied “classical” digital image and signal processing and my research focused on improving biomedical imaging technologies. At the time, neural networks were an interesting and novel field of research, but they had limited application (in image processing). When I graduated, the landmark AlexNet results had just been published and that is when the practical applications of deep learning really started to explode. I followed along with developments in convolutional neural networks over the next few years, but NeuroSoph was my first opportunity to dive into AI research full-time and to explore other areas of machine learning, like natural language processing.

What are the main challenges in your current role and why do they exist?

With respect to AI, the biggest challenge is simply staying up to date with the latest innovations and determining the most relevant and helpful information to share with the rest of our team. This research field is so dynamic that the amount of informational material published every day can be overwhelming at times. The great thing is that there are so many terrific resources available today – from traditional journal articles to online courses and web blogs, to corporate and academic conferences like this RE•WORK AI for Government Summit – that anyone with enough interest and time can learn all about the history, current state, and the potential future of AI. While it can be a challenge to keep pace with the latest developments, it’s also the most interesting part of my job.

How have recent progressions in AI helped advance your work and where have you seen the greatest improvements?

NeuroSoph has developed an automated intelligent forms processing system we named Specto, which is a customizable solution capable of classifying and extracting data from digitized forms. Advances in image recognition and classification algorithms, combined with continually improving NLP models, have enabled us to extract and interpret data with a high accuracy that was not achievable just a few years ago. Additionally, the proliferation of open source deep learning software libraries and frameworks (like TensorFlow, PyTorch, Caffe, MXNet, and many others) has enabled small companies like ours to efficiently devise, test, and deploy AI solutions to a variety of unique problems.

Which other industries are you most interested to see benefiting from AI in the next 5 years?

Given the current rate of development and growth of AI technologies, it’s difficult to imagine any industry that won’t benefit from AI in some way in the next five years. As a company, NeuroSoph is primarily focused on improving workflows in the government sector, where a lot of repetitive processes are still performed manually. So, government (at all levels) is the easy answer. Beyond that, I am particularly interested in how AI can augment and improve the different facets of healthcare – from diagnostic tools for physicians and patients, to efficient and secure information sharing between providers. We are already starting to see AI-powered technologies in use in the healthcare sector, and while careful attention must be given to ensure they are used appropriately, I am hopeful that they will ultimately improve and streamline a complicated healthcare system.

What are you most looking forward to at the AI for Government Summit, Toronto?

The NeuroSoph team and I are most looking forward to meeting and speaking with the many speakers and attendees, to hear about their experiences with AI in the government sector and to better understand the needs of different government agencies. I find conferences and summits like these to be the most informative and helpful resources for learning, because they bring together researchers, developers, and end-users, and facilitate productive discussions of current issues and state-of-the-art solutions.

If you want to learn more about the exciting work that NeuroSoph are doing, register for the AI for Government Summit here. Confirmed speakers for the summit include: MPP Hon. Peter Bethlenfalvy, President at the Treasury Board of Ontario, Ekkehard Ernst, Chief Macroeconomic Policy Unit at the International Labor Organization, Anton Prokopyev, Data Scientist at The World Bank and many more.