Community//

Women Leading The AI Industry: “ Keep calm and explore AI.” with Rosaria Silipo, Ph.D. and Tyler Gallagher

As previously stated, there are some open concerns, like the ethical choices that an AI machine must be forced to make, or people’s privacy. We definitely need to address them in the near future. In the meantime, though, let’s remain rational and not turn into technophobes. How can I say it? Keep calm and explore AI. […]


As previously stated, there are some open concerns, like the ethical choices that an AI machine must be forced to make, or people’s privacy. We definitely need to address them in the near future. In the meantime, though, let’s remain rational and not turn into technophobes. How can I say it? Keep calm and explore AI.

As part of my series about the women leading the Artificial Intelligence industry, I had the pleasure of interviewing Rosaria Silipo, Ph.D., principal data scientist at KNIME. Rosaria is the author of 50+ technical publications, including her most recent book “Practicing Data Science: A Collection of Case Studies.” She holds a doctorate degree in bioengineering and has spent more than 25 years working on data science projects for companies in a broad range of fields, including IoT, customer intelligence, the financial industry, and cybersecurity.


Thank you so much for doing this with us! Can you share with us the ‘backstory” of how you decided to pursue this career path?

As in most life paths, my career happened also a bit by chance. At the time of choosing the topic of my master thesis at the end of 1990, I decided on the automatic detection of patterns in the electrocardiographic signal, using statistical methods and neural networks. The topic of self-learning machines was fascinating. Compared to previous statistical methods, neural networks were stressing this self-learning feature. It is true that a neural network is eventually just a function whose parameters I can optimize, but its representation via neural units allows me a more pragmatic approach in shaping this function. It all started there. Since then, I have never really left the field of data analytics. There have been pauses of pure coding, but eventually, I always went back to the analysis of data. I have used data analysis on customer data, biomedical signals, automatic translations, speech recognition, banking, IoT, cybersecurity, bots, social media and on many more data domains. Looking back, it seems as if I really could never abandon that track. I tried but was always drawn back to data science.

What lessons can others learn from your story?

The lesson that can be learned is that a career stemming from math and algorithms is actually deeply enjoyable!

Can you tell our readers about the most interesting projects you are working on now?

Currently, I am working on a few “toy” projects to explore the creative side of AI. Using LSTM units within a simple neural network, I try to generate candidate names for new product lines or free texts in different languages, using different speaking styles. For English, for example, I generate Shakespeare-like text as well as rap songs. The network is the same; the training data are different. It is not just automation anymore. I am exploring the creative side of AI. One of my articles on using deep learning to write Shakespeare was recently published in InfoWorld: www.infoworld.com/article/3340377/deep-learning/how-to-use-deep-learning-to-write-shakespeare.html.

None of us are able to achieve success without some help along the way. Is there a particular person who you are grateful towards who helped get you to where you are? Can you share a story about that?

Oh man … There are so many people to thank for whatever it is that I do now! I could start with my mom … If I really have to restrict the number of people to thank to just a few, I would like to thank the KNIME founders, especially KNIME CEO Michael Berthold, for insisting on the concept of open source software and, therefore, giving a larger breadth to our work than the usual profit-earning objectives. I would also like to thank my former colleagues, especially the head of my group, Kate Knill at Nuance Communications, where I worked at the beginning of this millennium. There, I learned a new work ethic, relying on the enthusiasm for my job and on teamwork as a key to achieving greater goals.

What are the 5 things that most excite you about the AI industry? Why?

1. Creativity: Free Text Generation

One very exciting emerging topic is the application of AI to tasks that were traditionally deemed creative. So far, AI and data science have been successfully applied to predictive tasks and automatization of repetitive tasks. But we are starting now to see the potential of AI in more creative tasks, such as free text generation. Creative applications of AI can be limitless: from new fairy-tale text to bot more natural answers, from new product naming to song generation. This might lead to a second generation of AI applications that can tackle creativity better than the previous AI generation. In my opinion, this new set of applications is still in a somewhat experimental state, requiring a very large amount of computational resources and highly skilled human resources, which confines it for now to the field of research and experimentation. However, I can see great potential for rapid growth over the next few years.

Reference: //www.knime.com/blog/text-generation-with-lstm

2. AI-Powered New Journalism

We all know about the proliferation of fake and distorted content on social media and in the news. On the other hand, the mass of data to analyze is such that manual intervention by a few expert journalists cannot suffice anymore. In the future, the situation can only get worse with the amount of data that keeps growing. In the future, AI-powered journalism might help in the battle for the readers’ trust. Some projects have already started in an attempt to detect and label fake news or deceiving information content. These efforts can only grow together with the mass of data.

Reference: www.knime.com/blog/Forum_Authority_Sentiment

3. IoT Analysis

AI in IoT use cases is also becoming very popular. On one hand, IoT is an incredible source of data, and on the other hand, the analysis of IoT data produces quick monetization. AI in IoT use cases may vary in the future from simple demand prediction to anomaly detection in predictive maintenance, or from equipment monitoring to security attack detection.

Each one of these fields presents its own challenges and goals. However, all of them will need to deal with a large, constantly incoming flow of data as well as with new AI solutions. The problem of anomaly detection, for example, already requires a slightly different way of thinking than for traditional AI solutions. In the case that a class of anomalies is not available, different AI techniques based on uncertainty and the concept of the unknown are required.

Reference: https://files.knime.com/sites/default/files/181212_Whitepaper_Anomaly_Detection_Predictive_Maintenance_KNIME.pdf

4. Emergence of AI Integrations in Open Source Platforms

Another big challenge in the near future will be integration. With the steadily growing flow of data, integration of data formats, data sources, and even data analytics tools becomes the key for a successful approach to AI solutions.

In particular, I strongly believe that open source is the future, and I am excited about the recent emergence of AI frameworks integrated into existing open source environments (such as Python and KNIME). This will lead to the operationalization of AI techniques, allowing organizations to share, verify, refine, expand and reuse code as well as integrate data formats, data sources and other external tools. Looking at a five-year or longer vantage point, this may be the only way for an AI lab to keep up and remain relevant as the rate of technology advances accelerates.

5. Semi-supervised Learning for Labeling Large Amounts of Data

There is one kind of project that is showing up more and more often: the need to label all the collected data. In past years, companies have put a lot of efforts into digitalizing their data, processes and other information sources. Now, with all the data practically ready and piling up, the same companies are exploring ways to analyze it. The problem is that many AI techniques require a labeled subset of data. Labeling data is expensive and time-consuming. Thus, there will be growing attention devoted to those semi-supervised learning techniques, like active learning, that allow for quickly and inexpensively labeling a large subset of the original data.

What are the 5 things that concern you about the AI industry? Why?

1. Unreasonably High Expectations About What AI Can Do

The hype of AI/ML/DL may lead organizations to set unreasonable expectations about the results that can be obtained. I often read about the danger posed by a fully independent AI army to human kind. But when I train my AI models to learn to generate text, for example, I realize that expectations might be a bit too high about the status of AI development. It is possible that the current hype has already generated unrealistically high expectations and that this will produce a sense of disappointment in AI technology in the coming years.

2. AI Solutions Without AI Skills

You don’t need a Ph.D. to do data science; applications for fully automated machine learning and week-long training programs can get non-experts up and running. While this might work for simple and standard predictive analytics problems, organizations will still need access to data science experts to question suspiciously good results, find alternatives when something fails, look under the hood at the integrity of the algorithms, etc. The risk is to misuse AI algorithms which produce poor predictions and ultimately poor decision-making.

3. AI Governance Needed for Transparent AI

In the next few years, the integrity of AI might suffer in the rush to find the next best algorithm or the best predictions. Everyone is searching for the golden egg, and few are really paying attention to the integrity of the systems. Governance around the soundness of models and algorithms will be essential — someone needs to be looking under the hood to make sure that the data is accurate, the model/algorithms valid, etc. Benchmarks are few, but here is an interesting study that makes the point: https://onlinelibrary.wiley.com/doi/abs/10.1002/widm.1279.

4. Security Threats

The large collection of data and the large usage of data by AI technology might open a few doors to possible threats and attacks. This problem will become more and more prominent in the next few years and needs to be resolved quickly and effectively for people to trust their data to AI technologies.

5. Privacy and Ethics

And of course, last but not least, the problem of privacy and ethics. How far can data analysis go? Can the algorithm follow the single user and predict what he/she is going to do next, instant after instant? How far can the algorithm go into people’s privacy and history to formulate customized offers? And if the algorithm is wrong, who will take responsibility? This is, as of now, a gray area that will need to be resolved effectively and in a timely manner.

As you know, there is an ongoing debate between prominent scientists, (personified as a debate between Elon Musk and Mark Zuckerberg,) about whether advanced AI has the future potential to pose a danger to humanity. What is your position about this?

I just read an interesting interview with Garry Kasparov exactly on this topic. The title alone is enlightening: “We need better humans, not less technology.” There are some concerns, of course, about the new AI technology, but we cannot really burn the books it has been written on. Like electricity or the internet, humans need to get accustomed to the new technologies, to control them properly, and eventually to take them for granted. AI will be part of our future. How fast this will happen and how controlled this change will be depends on how we approach it now. It is naive to think that AI adoption can be avoided.

What can be done to prevent such concerns from materializing? And what can be done to assure the public that there is nothing to be concerned about?

As previously stated, there are some open concerns, like the ethical choices that an AI machine must be forced to make, or people’s privacy. We definitely need to address them in the near future. In the meantime, though, let’s remain rational and not turn into technophobes. How can I say it? Keep calm and explore AI.

How have you used your success to bring goodness to the world? Can you share a story?

Over the course of my career, 50+ technical papers, and thousands of tutorial videos, I have helped innumerable people to learn about data science, understand machine learning algorithms, create new higher education projects, and think creatively about data science applications. At a conference a few years ago, a young man approached me after my presentation. He wanted to thank me because he had grown his professional career by studying my tutorials, conference presentations, webinars and books. It made me feel old but proud of my work. Regardless of the scope of my influence, each person with whom I can share my knowledge, experience and enthusiasm is humbling and motivates me to do more.

As you know, there are not that many women in your industry. Can you share 3 things that you would you advise to other women in the AI space to thrive?

Besides studying a STEM discipline, I have some general advice.

1. Curiosity. Things change quickly in the field of data analytics. Even though statistics has been around for many years, new tailored data analytics algorithms, new tricks, and new techniques emerge every year. Keep an eye for innovative solutions on journals, blogs and reports. If something peaks your curiosity, run a small experiment to evaluate the expectations. Make sure you reserve a bit of time in your work to test some new technology from time to time. A small toy project is enough to get an idea of how useful the new technology can be and to get the creativity going.

2. Do not get discouraged. Sometimes customers have unrealistic expectations about the magic that data science can deliver. Sometimes, communications with the customer can go awry. Sometimes, you just need more time to provide a better solution. In all those cases, customers might be unhappy with the results. Do not get discouraged. Through this process, you have learned how to set expectations, run better communications, and the next useful steps for your next project.

3. Have fun at your job! We work for eight and more hours a day. If we do not enjoy it at least some of the time, it is going to eat all of our time with no return. All jobs make you deal with unpleasant moments, but make sure that at least some part of the job, as small as it may be, stimulates your curiosity, lets you learn something new, and unleashes your creativity.

Can you advise what is needed to engage more women into the AI industry?

The main problem in hiring women in our space is the lack of women graduating in STEM disciplines, especially in math, physics, engineering and computer science. I, personally, have managed to hire a number of women data scientists, but the candidate pool, I admit, is a bit limited. Machine learning techniques can be taught. However, a background in mathematics, statistics and computer science is still needed for a successful knowledge transfer.

What is your favorite “Life Lesson Quote”? Can you share a story of how that had relevance to your own life?

“The difference between theory and practice is that, in theory, they are the same.” I have used this well-recognized quote often, from setting expectations on a project to commenting on weird results, from discussing machine precision to programming errors, from buying myself more time (and patience) on a project to unexpected turns in my private life. It is just the motto I use to remain positive when things do not go as expected.

You are a person of great influence. If you could start a movement that would bring the most amount of good to the most amount of people, what would that be? You never know what your idea can trigger. 🙂

I would like to expand the adoption of open source software. Open source software is not only a great way to download free stuff. It is really a great professional philosophy that allows you to leave behind that horrible “customer = money” attitude and embrace a more community-oriented way of working. It makes your professional life much more satisfying — maybe not less busy but centered around the more valuable tasks!

How can our readers follow you on social media?

Follow Rosaria on Twitter, LinkedIn and the KNIME blog. For more information on KNIME, please visit www.knime.com.

This was very inspiring. Thank you so much for joining us!

The Thrive Global Community welcomes voices from many spheres. We publish pieces written by outside contributors with a wide range of opinions, which don’t necessarily reflect our own. Learn more or join us as a community member!
Share your comments below. Please read our commenting guidelines before posting. If you have a concern about a comment, report it here.

Sign up for the Thrive Global newsletter

Will be used in accordance with our privacy policy.

Thrive Global
People look for retreats for themselves, in the country, by the coast, or in the hills . . . There is nowhere that a person can find a more peaceful and trouble-free retreat than in his own mind. . . . So constantly give yourself this retreat, and renew yourself.

- MARCUS AURELIUS

We use cookies on our site to give you the best experience possible. By continuing to browse the site, you agree to this use. For more information on how we use cookies, see our Privacy Policy.