This is the translation (made by me, so any typos or funny-sounding bits are all mine!) of an interview I did with Media INAF in December 2020. Credits: Media INAF and Valentina Guglielmo.
The names and stories of 50 women who made - and still make - the history of information technology is the subject of a list published by Wired.it last month. Women of the past and present, of all nationalities and social background. Women who pursued higher education and worked hard to make a difference — like Stephanie Shirley, known by her colleagues as "Steve," a male nickname that allowed her to succeed and gain much-deserved career advancement in a male-dominated world. Women like Annie Easley, who co-developed a rocket launcher software, indispensable for the Shuttle and the Cassini mission. Women like the first woman with a Ph.D. in computer science, the CEO of YouTube, the developer of the renowned Apple graphics, and Facebook's COO. Five Italians and two astrophysicists are on the list; at the intersection of the two categories we find Viviana Acquaviva. Originally from Lecce, in the Apulia region of Italy (t/n), she holds a Master's degree in Physics and a Ph.D. in Astrophysics from the International School for Advanced Studies in Trieste, she has worked for fourteen years in the United States and for nine years in New York. She is an Associate Professor in the Physics department of the CUNY NYC College of Technology and at the CUNY Graduate Center. She is currently writing a textbook on machine learning techniques applied to Physics and Astrophysics research, to be published by Princeton University Press. Media INAF joined her in New York, at the end of her sabbatical year in Barcelona.
You have lived and worked in the United States for many years; last year you returned to Europe. Do you miss Italy at this moment, in which you are basically forced to stay away?
Every now and then I entertain dreams of returning to Italy, but I know that in the near future it will be difficult. You know, when you are away you always live with a split heart; for example, my husband is Egyptian, and we met here, and part of his family is also here. I certainly miss Europe; despite the fact that many years have passed, many of my closest friendships are with Europeans and Italians. Academia however is a peculiar world, the professorship in a university is not easily transferable, and New York is a special place, in terms of opportunities; I like my job very much and I wouldn't change it. So, as long as we are in a phase in which work counts, makes us feel fulfilled, and gives us the chance to do beautiful and interesting things, I am satisfied with spending my vacation time in Italy, and I accept living with a slightly broken heart.
You were listed by Wired.it among the fifty women who have made (and make) the history of information technology. You are among very important figures, to whom we owe the birth of computer science, or who have created many of the systems, codes and technologies we use today. How do you feel about being on that list?
I must admit that I wondered: how did they assemble it? I like my job, I like teaching and I like doing it in a public university here in the US. We recently started a new major in Physics and I, as one of the organizers, pushed hard to include courses in applied computational Physics. I graduated in 2002, and I have often witnessed, in the academic world, the idea that if you major in Physics, then you have to do your doctorate, you have to spend your days doing incomprehensible calculations at the blackboard. I think it is an outdated idea, particularly so here, in the United States, where there is a very wide social gap between those attending public and private schools: the kids who graduate from my university need to find a good job, not necessarily to be encouraged to undertake a path similar to my own. They need the ability to move from a job they do, while studying, just to pay their bills, to one that leads to a fulfilling career. Because of this, I insisted on including machine learning courses, which is my field, and also quantum computing and advanced computational methods, so that an employer will see useful practical skills in their resume, not just theory. I would say this is my claim to fame. How Wired.it came to learn about this, however, I have no idea. Two years ago I was named one of "Inspiring Fifty", a collection of 50 Italian women with some impact in the world of technology, created by a non-profit organization with the goal of increasing female representation in this field. How do you end up on these lists in first place? I think it's serendipitous. They may have looked for profiles and stories that looked interesting. I gladly accept this recognition and I am not ashamed of it; I'll be happy if someone feels inspired by seeing it or just thinks it is worth a read. But I don't consider myself comparable to Joan Clarke, or Grace Hopper.
So, you don't think that your approach to scientific research and teaching can "make history" - even if only in the sense of changing the way of doing research or changing the vision towards a course of study such as Physics?
On the contrary, I think so. I believe in my attempt to change the traditional idea of the physicist and propose one that is a little more modern and functional, and I also hope that this will help some women who do not feel up to such a career. But I am also aware that I play in a small field - that my impact won't be huge. Who knows, maybe at some point there will be five or ten people who will change their mind, and when I reach the end of my career these people may be one hundred and that's enough - indeed, this is one of the reasons why I like my job.
What is your idea of the modern physicist or astrophysicist?
My main idea is that everyone has to find out what works for them. I work in the field of Astrophysics but I have never been a child who loved watching the stars. I thought I would be doing math until just before enrolling in university, then I studied Physics, not Astronomy. When the time came to decide on the topic for my doctorate, at the International School for Advanced Studies in Trieste, I had a choice between astrophysics and particle physics. I met a professor, Carlo Baccigalupi, and I really liked his field of research, but above all I liked him as a person. I didn't have a linear path and I struggled for many years to accept the legitimacy of not being born with a passion for Astronomy. I would say that I like mostly methods and problems, I like to mess around with computer programs, and I could be a marine biologist just as like as a physicist - I think I would be equally happy. This is to say that my first suggestion is to abandon the idea that there is a predefined path to follow. There must be no preconceptions, and our duty as instructors is to provide skills that open doors to more opportunities, and to foster students' skills in scientific reasoning and analytical thinking, as well as programming. As advisors, we need to get rid of the idea that if students do not pursue an academic career it is a failure, and ask them instead what they want to do. So, I don't have a precise idea of how a physicist should be, but: not necessarily a man, not necessarily white (t/n), not necessarily old, not necessarily at the blackboard. These are non-negotiable points.
The immediate connection you made between a Physics degree and the job market is interesting. Above all, it is unusual, given that when students enroll in Physics, it is often assumed that they are driven by the desire to do research. This is why we are often surprised - and a little disappointed, perhaps - when this is not the case.
It's so true. I must say that, in my reasoning, there is also the awareness of a fundamental difference between Europe and the United States. In Europe - and in Italy - some basic rights are guaranteed, and it is possible to live without making salary the very first thing to look at. Having the right to health care, the right to education for your children, and so on, makes it still possible to think that, if you like doing research, a doctorate is the way to go. Even so, I don't think it is right to always encourage this career path, because if we compare the number of permanent positions available in research and the number of graduates, the bottleneck is obvious.
How about in the US?
In the US this aspect is exacerbated by the fact that some fundamental rights do not exist. My students in particular, who come from public school, do not have rich families behind them to give them financial support, and most often have to work to support their studies in college. We don't necessarily change their life by encouraging them to spend a thousand dollars to make ten applications for graduate programs; maybe we can also explain that they can go to work immediately and that they can also have a good career in data science.
I had a very good student, who worked with me for two years; he had also won a competitive scholarship to support students who want to pursue their doctorate. Then, as soon as he graduated, he received a full-time apprenticeship offer from Amazon. He came to me almost in tears thinking that I might feel disappointed, telling me that he was no longer sure of his intentions. I asked him what he wanted to do, he replied that by accepting that job he could help his brother - who paid for his studies by driving a taxi - to finish high school, he could buy a bigger house for his mother, and they could all live together. I told him, if the passion for research remains, nothing prevents you from thinking about it and returning to graduate school in a few years. He accepted the job, and after only a year and a half he succeeded in all the goals he had set for himself. How can we blindly push for an academic career, when we face situations like this every day? It would already be wrong as a general rule, in my opinion; in the context of my job, it becomes really short-sighted advice. We need to understand the needs of our students, and there is still a lot of work to do on this.
Wired.it writes that you are writing a book. What is it about?
It's a textbook. At CUNY, I teach a Machine Learning for Physics and Astronomy course; I've taught it five times. The methods we use are the same that are found in many books and online courses - I myself have taken many of those classes, the last one last year. When addressing these courses from a computer science perspective, however, the approach is very technical. Instead, if you do science, the most important thing is to understand how to integrate the scientific method with these techniques: to understand which variables are important, and to develop the ability to decide the right method to tackle a certain type of problem. This type of content is hard to find in the online courses. I put together the material based on the experience I accumulated during teaching, and by collecting a series of physics-based worked examples, similar to those we encounter every day in research. I was thinking that it would have been very convenient for me to have a textbook, when I was learning the ropes. Last year I was able to take a sabbatical year to work on the book, which we spent - with my husband and my daughter - in Barcelona, and the idea of this book was very well received by Princeton University Press, which became my publisher.
Who is the main target, your students?
You know, students are young and resilient, they have fewer obstacles in learning, even if they have to search for and assemble material from different sources. I think also of the professors who teach in Physics and Astronomy departments, who will have to - if they have not already done so - begin to approach this subject and teach machine learning in their courses, because it is becoming increasingly important. Maybe for them it will be helpful to have a fairly complete, single source like this textbook, which includes notebooks, review questions, slides and exams, and will also be accompanied by a series of practical examples structured as small research projects based on public data. I hope it will facilitate the process to make it an approachable subject - even if we (instructors) have not studied it in our own course of study. Similarly, I hope it can also be useful to Ph.D. students and postdocs, who want to learn new methods for their research on their own.
When will your book come out?
I spoke to my publisher just yesterday, I will finish writing the first draft in February and I hope it will be published in about a year. I also hope it becomes a living document, which I will continue to update (t/n).
What are the limits of machine learning - if any - in your opinion?
I would say there are two, one more important than the other. First of all, machine learning is not magical. In my day, we studied from a big book called Numerical Recipes - with a lot of linear algebra - and I think of machine learning as its modern-day evolution. It is a set of tools, and as such the worst danger is the fact that it can be applied in a blind, obtuse way, affecting the quality of the science that is produced with it. It is a real risk, because these are methods that are perhaps a little more complicated, less transparent than analytical models, for example, and therefore more subject to wrong interpretations. In the last five or ten years, there has been a rush to publish articles that employ machine learning, and it was very easy for them to go through the peer review process, perhaps as a consequence of our limited ability to critique them. We may have been a bit hasty and easy-going in publishing, while we must have more awareness and be more cautious: we need to examine machine learning works with the same rigor that we put in reviewing any scientific paper, and make a greater effort in understanding and evaluating the methods used.
The second one?
The second concerns the so-called “ethical biases”, and it is much more important, because it concerns society as a whole and not the small community of (astro)physicists. On the one hand, many worthwhile goals can be achieved with machine learning - such as early detection of diseases, or the study of rare diseases, or the process of making resources, like water, electricity, energy, accessible and of optimizing their distribution. This of course relates to the very serious issue of climate change. On the other hand, there is the potential for greatly dangerous applications. I'm thinking about deepfakes, or about passing on and increasing race and gender biases - all of which will happen, in the absence of precise regulations. Once again, if Europe tends to have a slightly tighter hand, it is not the same in the United States. And since the world of technology and information technology moves very quickly, it is difficult for regulations to keep up.
Can you give a positive example?
Rare diseases. If a child is born with a rare genetic condition, it used to be unlikely that they would be cured - nobody wants to invest in treating three people in the world. But if you can study how such a rare combination connects, for example, to another similar strain, and then another and another - and that's what machine learning does - then suddenly this disease becomes a notable thing, and it will be deemed to be worth studying. So, I don't want to give in to pessimism, but I think it is very important to build rules and that it is up to us practitioners - and there are still too few of us - to get involved and teach these things.
Do you think that there are job categories at risk, in relation to the spread of machine learning techniques?
That's a good question. I can only answer for what concerns the world of research, because I feel I can give a somewhat knowledgeable answer: I don't think it's a bad thing. The tasks we automate with machine learning are typically long and boring - we will never replace creative thinking. To give an example, you used to have to look at ten thousand images to search for a supernova, or you would have to spend days checking catalogs to look for possible contaminants. I would say that if we automate these operations we do not lose anything; these are not activities that people find fulfilling. Therefore, in research I would say that this is not an issue, and it is still up to the researcher to decide which method to use and how to implement it. Outside the world of research, however, I would not know. I don't want to give a superficial answer, but I think it is possible that some professional roles can and will be replaced - as it happened with the industrial revolution. And I think this can be a risk, especially for those who find it difficult to reinvent themselves and find a new job, more so if they are towards end of their career. I don't know what the balance could be between these and the new professions that are created - however, I realize that those won't be easily accessible to the same workers who see their work diminished.
Going back to research again, do you think that the spread of new and brilliant career prospects related to machine learning can take minds away from research?
Maybe. But it is a welcome effect; I see it as a very positive thing. I see it as an opportunity for the academic world to realize that researchers, especially those on temporary contracts, need to be treated better. They are very smart people, with advanced and hard-to-find skills, and they can have a wonderful career outside of science. I think it's a limitation of our academic culture to think that if you like your job, then it's not a job. The hours spent on a project, on a proposal, cost effort and time. Researchers must realize that their time is valuable, and be given the means to progress and undertake profitable and satisfying careers, which allow a good work-life balance. This is an aspect on which academia has a lot to learn.
You were telling me that you have a four-year-old daughter. Do you also teach her the correct way to approach information technology and learn how to use it?
It is very nice to see children learn. I received a very traditional education: my mom is a mathematician, just like my grandmother, and every meal at the table was accompanied by math questions - how could you calculate this or that? I was a child who liked to learn and study, and I liked my time alone. My daughter is much more dynamic. When the lockdown occurred, we were in Barcelona, we spent many weeks indoors. She had just turned four, and she became very interested in reading and writing: she would go around the house with this little board, writing and reading words all day. Once we were able to go out again, she quit and now she has zero interest for reading - I assume she's learning other things. My challenge with her is to not focus on what worked for me, on the things I liked to do, and to encourage her natural curiosity instead. But one thing I do regularly is to read many stories for her - she likes to hear about successful women in science a lot - and she is finding her idols and examples as well. The only message I always try to give her is that she can do what she wants, that she must not be feel restricted or limited because of certain ideas or preconceptions. She seems like a happy child to me, I can't ask for more than this.
The names and stories of 50 women who made - and still make - the history of information technology is the subject of a list published by Wired.it last month. Women of the past and present, of all nationalities and social background. Women who pursued higher education and worked hard to make a difference — like Stephanie Shirley, known by her colleagues as "Steve," a male nickname that allowed her to succeed and gain much-deserved career advancement in a male-dominated world. Women like Annie Easley, who co-developed a rocket launcher software, indispensable for the Shuttle and the Cassini mission. Women like the first woman with a Ph.D. in computer science, the CEO of YouTube, the developer of the renowned Apple graphics, and Facebook's COO. Five Italians and two astrophysicists are on the list; at the intersection of the two categories we find Viviana Acquaviva. Originally from Lecce, in the Apulia region of Italy (t/n), she holds a Master's degree in Physics and a Ph.D. in Astrophysics from the International School for Advanced Studies in Trieste, she has worked for fourteen years in the United States and for nine years in New York. She is an Associate Professor in the Physics department of the CUNY NYC College of Technology and at the CUNY Graduate Center. She is currently writing a textbook on machine learning techniques applied to Physics and Astrophysics research, to be published by Princeton University Press. Media INAF joined her in New York, at the end of her sabbatical year in Barcelona.
You have lived and worked in the United States for many years; last year you returned to Europe. Do you miss Italy at this moment, in which you are basically forced to stay away?
Every now and then I entertain dreams of returning to Italy, but I know that in the near future it will be difficult. You know, when you are away you always live with a split heart; for example, my husband is Egyptian, and we met here, and part of his family is also here. I certainly miss Europe; despite the fact that many years have passed, many of my closest friendships are with Europeans and Italians. Academia however is a peculiar world, the professorship in a university is not easily transferable, and New York is a special place, in terms of opportunities; I like my job very much and I wouldn't change it. So, as long as we are in a phase in which work counts, makes us feel fulfilled, and gives us the chance to do beautiful and interesting things, I am satisfied with spending my vacation time in Italy, and I accept living with a slightly broken heart.
You were listed by Wired.it among the fifty women who have made (and make) the history of information technology. You are among very important figures, to whom we owe the birth of computer science, or who have created many of the systems, codes and technologies we use today. How do you feel about being on that list?
I must admit that I wondered: how did they assemble it? I like my job, I like teaching and I like doing it in a public university here in the US. We recently started a new major in Physics and I, as one of the organizers, pushed hard to include courses in applied computational Physics. I graduated in 2002, and I have often witnessed, in the academic world, the idea that if you major in Physics, then you have to do your doctorate, you have to spend your days doing incomprehensible calculations at the blackboard. I think it is an outdated idea, particularly so here, in the United States, where there is a very wide social gap between those attending public and private schools: the kids who graduate from my university need to find a good job, not necessarily to be encouraged to undertake a path similar to my own. They need the ability to move from a job they do, while studying, just to pay their bills, to one that leads to a fulfilling career. Because of this, I insisted on including machine learning courses, which is my field, and also quantum computing and advanced computational methods, so that an employer will see useful practical skills in their resume, not just theory. I would say this is my claim to fame. How Wired.it came to learn about this, however, I have no idea. Two years ago I was named one of "Inspiring Fifty", a collection of 50 Italian women with some impact in the world of technology, created by a non-profit organization with the goal of increasing female representation in this field. How do you end up on these lists in first place? I think it's serendipitous. They may have looked for profiles and stories that looked interesting. I gladly accept this recognition and I am not ashamed of it; I'll be happy if someone feels inspired by seeing it or just thinks it is worth a read. But I don't consider myself comparable to Joan Clarke, or Grace Hopper.
So, you don't think that your approach to scientific research and teaching can "make history" - even if only in the sense of changing the way of doing research or changing the vision towards a course of study such as Physics?
On the contrary, I think so. I believe in my attempt to change the traditional idea of the physicist and propose one that is a little more modern and functional, and I also hope that this will help some women who do not feel up to such a career. But I am also aware that I play in a small field - that my impact won't be huge. Who knows, maybe at some point there will be five or ten people who will change their mind, and when I reach the end of my career these people may be one hundred and that's enough - indeed, this is one of the reasons why I like my job.
What is your idea of the modern physicist or astrophysicist?
My main idea is that everyone has to find out what works for them. I work in the field of Astrophysics but I have never been a child who loved watching the stars. I thought I would be doing math until just before enrolling in university, then I studied Physics, not Astronomy. When the time came to decide on the topic for my doctorate, at the International School for Advanced Studies in Trieste, I had a choice between astrophysics and particle physics. I met a professor, Carlo Baccigalupi, and I really liked his field of research, but above all I liked him as a person. I didn't have a linear path and I struggled for many years to accept the legitimacy of not being born with a passion for Astronomy. I would say that I like mostly methods and problems, I like to mess around with computer programs, and I could be a marine biologist just as like as a physicist - I think I would be equally happy. This is to say that my first suggestion is to abandon the idea that there is a predefined path to follow. There must be no preconceptions, and our duty as instructors is to provide skills that open doors to more opportunities, and to foster students' skills in scientific reasoning and analytical thinking, as well as programming. As advisors, we need to get rid of the idea that if students do not pursue an academic career it is a failure, and ask them instead what they want to do. So, I don't have a precise idea of how a physicist should be, but: not necessarily a man, not necessarily white (t/n), not necessarily old, not necessarily at the blackboard. These are non-negotiable points.
The immediate connection you made between a Physics degree and the job market is interesting. Above all, it is unusual, given that when students enroll in Physics, it is often assumed that they are driven by the desire to do research. This is why we are often surprised - and a little disappointed, perhaps - when this is not the case.
It's so true. I must say that, in my reasoning, there is also the awareness of a fundamental difference between Europe and the United States. In Europe - and in Italy - some basic rights are guaranteed, and it is possible to live without making salary the very first thing to look at. Having the right to health care, the right to education for your children, and so on, makes it still possible to think that, if you like doing research, a doctorate is the way to go. Even so, I don't think it is right to always encourage this career path, because if we compare the number of permanent positions available in research and the number of graduates, the bottleneck is obvious.
How about in the US?
In the US this aspect is exacerbated by the fact that some fundamental rights do not exist. My students in particular, who come from public school, do not have rich families behind them to give them financial support, and most often have to work to support their studies in college. We don't necessarily change their life by encouraging them to spend a thousand dollars to make ten applications for graduate programs; maybe we can also explain that they can go to work immediately and that they can also have a good career in data science.
I had a very good student, who worked with me for two years; he had also won a competitive scholarship to support students who want to pursue their doctorate. Then, as soon as he graduated, he received a full-time apprenticeship offer from Amazon. He came to me almost in tears thinking that I might feel disappointed, telling me that he was no longer sure of his intentions. I asked him what he wanted to do, he replied that by accepting that job he could help his brother - who paid for his studies by driving a taxi - to finish high school, he could buy a bigger house for his mother, and they could all live together. I told him, if the passion for research remains, nothing prevents you from thinking about it and returning to graduate school in a few years. He accepted the job, and after only a year and a half he succeeded in all the goals he had set for himself. How can we blindly push for an academic career, when we face situations like this every day? It would already be wrong as a general rule, in my opinion; in the context of my job, it becomes really short-sighted advice. We need to understand the needs of our students, and there is still a lot of work to do on this.
Wired.it writes that you are writing a book. What is it about?
It's a textbook. At CUNY, I teach a Machine Learning for Physics and Astronomy course; I've taught it five times. The methods we use are the same that are found in many books and online courses - I myself have taken many of those classes, the last one last year. When addressing these courses from a computer science perspective, however, the approach is very technical. Instead, if you do science, the most important thing is to understand how to integrate the scientific method with these techniques: to understand which variables are important, and to develop the ability to decide the right method to tackle a certain type of problem. This type of content is hard to find in the online courses. I put together the material based on the experience I accumulated during teaching, and by collecting a series of physics-based worked examples, similar to those we encounter every day in research. I was thinking that it would have been very convenient for me to have a textbook, when I was learning the ropes. Last year I was able to take a sabbatical year to work on the book, which we spent - with my husband and my daughter - in Barcelona, and the idea of this book was very well received by Princeton University Press, which became my publisher.
Who is the main target, your students?
You know, students are young and resilient, they have fewer obstacles in learning, even if they have to search for and assemble material from different sources. I think also of the professors who teach in Physics and Astronomy departments, who will have to - if they have not already done so - begin to approach this subject and teach machine learning in their courses, because it is becoming increasingly important. Maybe for them it will be helpful to have a fairly complete, single source like this textbook, which includes notebooks, review questions, slides and exams, and will also be accompanied by a series of practical examples structured as small research projects based on public data. I hope it will facilitate the process to make it an approachable subject - even if we (instructors) have not studied it in our own course of study. Similarly, I hope it can also be useful to Ph.D. students and postdocs, who want to learn new methods for their research on their own.
When will your book come out?
I spoke to my publisher just yesterday, I will finish writing the first draft in February and I hope it will be published in about a year. I also hope it becomes a living document, which I will continue to update (t/n).
What are the limits of machine learning - if any - in your opinion?
I would say there are two, one more important than the other. First of all, machine learning is not magical. In my day, we studied from a big book called Numerical Recipes - with a lot of linear algebra - and I think of machine learning as its modern-day evolution. It is a set of tools, and as such the worst danger is the fact that it can be applied in a blind, obtuse way, affecting the quality of the science that is produced with it. It is a real risk, because these are methods that are perhaps a little more complicated, less transparent than analytical models, for example, and therefore more subject to wrong interpretations. In the last five or ten years, there has been a rush to publish articles that employ machine learning, and it was very easy for them to go through the peer review process, perhaps as a consequence of our limited ability to critique them. We may have been a bit hasty and easy-going in publishing, while we must have more awareness and be more cautious: we need to examine machine learning works with the same rigor that we put in reviewing any scientific paper, and make a greater effort in understanding and evaluating the methods used.
The second one?
The second concerns the so-called “ethical biases”, and it is much more important, because it concerns society as a whole and not the small community of (astro)physicists. On the one hand, many worthwhile goals can be achieved with machine learning - such as early detection of diseases, or the study of rare diseases, or the process of making resources, like water, electricity, energy, accessible and of optimizing their distribution. This of course relates to the very serious issue of climate change. On the other hand, there is the potential for greatly dangerous applications. I'm thinking about deepfakes, or about passing on and increasing race and gender biases - all of which will happen, in the absence of precise regulations. Once again, if Europe tends to have a slightly tighter hand, it is not the same in the United States. And since the world of technology and information technology moves very quickly, it is difficult for regulations to keep up.
Can you give a positive example?
Rare diseases. If a child is born with a rare genetic condition, it used to be unlikely that they would be cured - nobody wants to invest in treating three people in the world. But if you can study how such a rare combination connects, for example, to another similar strain, and then another and another - and that's what machine learning does - then suddenly this disease becomes a notable thing, and it will be deemed to be worth studying. So, I don't want to give in to pessimism, but I think it is very important to build rules and that it is up to us practitioners - and there are still too few of us - to get involved and teach these things.
Do you think that there are job categories at risk, in relation to the spread of machine learning techniques?
That's a good question. I can only answer for what concerns the world of research, because I feel I can give a somewhat knowledgeable answer: I don't think it's a bad thing. The tasks we automate with machine learning are typically long and boring - we will never replace creative thinking. To give an example, you used to have to look at ten thousand images to search for a supernova, or you would have to spend days checking catalogs to look for possible contaminants. I would say that if we automate these operations we do not lose anything; these are not activities that people find fulfilling. Therefore, in research I would say that this is not an issue, and it is still up to the researcher to decide which method to use and how to implement it. Outside the world of research, however, I would not know. I don't want to give a superficial answer, but I think it is possible that some professional roles can and will be replaced - as it happened with the industrial revolution. And I think this can be a risk, especially for those who find it difficult to reinvent themselves and find a new job, more so if they are towards end of their career. I don't know what the balance could be between these and the new professions that are created - however, I realize that those won't be easily accessible to the same workers who see their work diminished.
Going back to research again, do you think that the spread of new and brilliant career prospects related to machine learning can take minds away from research?
Maybe. But it is a welcome effect; I see it as a very positive thing. I see it as an opportunity for the academic world to realize that researchers, especially those on temporary contracts, need to be treated better. They are very smart people, with advanced and hard-to-find skills, and they can have a wonderful career outside of science. I think it's a limitation of our academic culture to think that if you like your job, then it's not a job. The hours spent on a project, on a proposal, cost effort and time. Researchers must realize that their time is valuable, and be given the means to progress and undertake profitable and satisfying careers, which allow a good work-life balance. This is an aspect on which academia has a lot to learn.
You were telling me that you have a four-year-old daughter. Do you also teach her the correct way to approach information technology and learn how to use it?
It is very nice to see children learn. I received a very traditional education: my mom is a mathematician, just like my grandmother, and every meal at the table was accompanied by math questions - how could you calculate this or that? I was a child who liked to learn and study, and I liked my time alone. My daughter is much more dynamic. When the lockdown occurred, we were in Barcelona, we spent many weeks indoors. She had just turned four, and she became very interested in reading and writing: she would go around the house with this little board, writing and reading words all day. Once we were able to go out again, she quit and now she has zero interest for reading - I assume she's learning other things. My challenge with her is to not focus on what worked for me, on the things I liked to do, and to encourage her natural curiosity instead. But one thing I do regularly is to read many stories for her - she likes to hear about successful women in science a lot - and she is finding her idols and examples as well. The only message I always try to give her is that she can do what she wants, that she must not be feel restricted or limited because of certain ideas or preconceptions. She seems like a happy child to me, I can't ask for more than this.