Coffee Chat with Tobias Gerstenberg

Today we hosted a coffee chat with new psychology professor Tobias Gerstenberg, who studies casuality in cognition, as well as counterfactuals and responsibility. Gerstenberg started off by detailing his journey to where he is now. As a native German, he initially attended university in Berlin, but later completed a MS and PhD at University College London. He then spent five years as a postdoc at MIT in a computational cognitive science group. 

Gerstenberg is excited to be at Stanford due to the cross-departmental collaborations that occur here, or, in some sense, the existence of SymSys at the faculty level. He interested in intersecting social psychology and AI with his own research and taking part in Stanford’s new Human-Centered Artificial Intelligence initiative. 

As the discussion about causality went on, Gerstenberg showed students the Heider-Simmel animation as an example of how humans infer social variables and attribute causality in very simple stimuli. Towards the end he also showed some examples of stimuli from his own research on intuitive physics, where study participants observe a falling ball hitting obstacles in a box. 

Written by Lucy Li

Humans of SymSys: Zach Harned

“SymSys offered me a chance to gain experience and specialization in a few important cognate disciplines.”

Zach Harned is a new SymSys master’s student who is also a J.D. Candidate at Stanford Law School. He is the founder of the Stanford Artificial Intelligence & Law Society (SAILS).

What drew you to the SymSys MS program as a law student? 

Much of the legal/policy/regulatory work in emerging technology requires a multi-disciplinary perspective. SymSys offered me a chance to gain experience and specialization in a few important cognate disciplines. 

What's something cool you've worked on at Stanford? 

One of my favorite Stanford projects is the Stanford Artificial Intelligence & Law Society (SAILS), which I founded in November of 2017. Running this organization has been incredibly fun. We created a lecture series where we brought in speakers working at the forefront of legal or policy issues related to artificial intelligence. We were lucky to find numerous stakeholders to help fund our organization, including Law, Science, and Technology, CodeX, and SPICE. 

What kind of research are you hoping to do in the MS program?

I've been focusing my research broadly on the legal implications of machine learning and artificial intelligence. This is a large and expanding area, so I've been attempting to circumscribe my interest into two main areas. First is FAT/ML, which stands for Fairness, Accountability, and Transparency in Machine Learning. The second area is machine learning in healthcare. In these two areas, there are a variety of novel issues, technologies, and techniques continually developing, all of which have a high impact on society. I've submitted a few papers for publication in these areas, that hopefully will be released soon. I was very fortunate to have fantastic collaborators from computer science and medicine to co-author these articles with me. 

What SymSys-related topic are you excited about right now?

Right now I'm heavily focused on how machine vision applications in healthcare will impact legal liability and medical malpractice. This is a fascinating problem, with many distinct players, complicated and cutting-edge technology, and little precedent. I am investigating how this technology impacts physicians themselves, hospitals as organizations, and software or medical device manufacturers.  

Humans of SymSys: Santosh Murugan

“I took a piece of paper, and wrote down a list of domains that I thought, if there was to be a major breakthrough, would make future generations of humans live in a far better world.”

Santosh is a SymSys senior and Advising Fellow who helps prospective and current students navigate the major. His office hours are Tuesday 3-4pm, Wednesday 2:00-2:55pm, and Thursday 3-5pm in 460-040A.

What drew you to SymSys? What’s your current concentration?

SymSys deals with the fundamental building blocks of what it means to be human: how we think (neuroscience), how we behave (psychology), how we speak (linguistics), how we augment our capabilities (computer science), and why we do these things (philosophy). After deliberating between concentrating in Artificial Intelligence and Human-Computer Interaction, I chose to pursue HCI first, with the intention of studying AI afterwards (prior to med school).

What’s your favorite SymSys-related class that you’ve taken?

In Spring 2018, I took MED 275: Biodesign Fundamentals to fulfill a requirement in the Human-Computer Interaction concentration. I came into the class with tempered expectations (the medical-device development pathway is notoriously brutal), but was absolutely blown away by the quality of instruction. In that class, I met the eventual co-founder for my biotech startup (read below!), met some incredible medical experts and mentors, and came to really appreciate some of the unique experiences that Stanford offers.

As an AF, What's one piece of advice you'd like to offer to younger students?

One of my favorite athletes, Wayne Gretzky, once said “I skate to where the puck is going to be, not where it has been.”

The truth is, within our lifetimes, the world is going to look a lot different than it does currently. The disruption has already started - think about what Amazon has done to the retail industry, or Tesla to the energy/automotive markets, in the last five years alone. The level of disruption effected by these companies is almost unfathomable.

But that’s just the tip of the iceberg. Both NASA and SpaceX are planning for multi-planetary habitation within the next twenty years. Biotechnologies like CRISPR and CAR-T are literally pushing the boundaries of science and human health. Artificial Intelligence (AI) has the potential to improve nearly every industry you can think of. I was in China last month, and saw literally hundreds of posters for AI conferences in street corners and on lamp posts.

I often see younger students struggling to choose what field to go into, and I feel for them- I know what that’s like. I can’t tell you exactly what will work for you, but I can tell you what helped me clarify some of my goals.

Inspired by a speech from a major tech luminary - I took a piece of paper, and wrote down a list of domains that I thought, if there was to be a major breakthrough, would make future generations of humans live in a far better world. Then, I picked the three that seemed most pressing and interesting to me (AI, biotech, and space). Once I had these clear goals (eliminating the paralysis of choice), it became much easier to focus on what was important to me, and to get to work on implementing these visions.

Once you pick your goals, seek advice from mentors, course-correct when necessary, and work diligently. I can’t guarantee that everything will work out, but at least you can rest happily knowing that you’re working to make the world a little brighter for the people around you.

What's something cool you've worked on?

The summer after my freshman year at Stanford, I had the opportunity to work computational genomics/oncology at the Dana Farber Cancer Institute, and the Broad Institute of MIT and Harvard in Boston. As part of this work, I shadowed in one of the world’s foremost neuro-oncology clinics at Massachusetts General Hospital. In my time at MGH, I met some of the most incredible patients - people whose entire lives had been upended by their cancer diagnosis, and yet remained some of the most incredibly resilient and inspiring people I’ve ever met. Simultaneously, on the research side, I worked with MDs and PhDs (e.g. Scott Carter and Priscilla Brastianos), who had dedicated their lives to give these patients a second chance, by inventing bio-technologies and designing clinical trials to leverage this technology and deliver it directly to patients. In that summer, I became increasingly inspired to build biomedical technologies that could play a huge role in changing people’s lives for the better.

That inspiration continues to this day. After realizing how devastating chemotherapy-induced hair loss can be for cancer patients, I co-founded a biotech startup dedicated to helping cancer patients keep their hair during and after chemotherapy. That startup graduated from Stanford’s Biodesign NEXT (incubator-style) program, and is continuing to make progress as part of the Cardinal Ventures accelerator. Simultaneously, in my work at NASA this past summer, I went heavy into genetic engineering (in this case, for building multiplanetary habitats via genetically engineered microorganisms) - and hope to apply the same fundamental techniques and knowledge to cure genetics-based human medical problems.

What (loosely) SymSys-related topic are you excited about right now?

Artificial Intelligence is pretty tightly-related to SymSys (it’s actually one of the concentrations), but it’s definitely one of the most exciting topics for me, personally. The fundamental improvements we’ve seen in a variety of fields (e.g. in medicine) as a result of machine learning/AI are mind-blowing; and although there are a lot of issues to be resolved, I’m hopeful that we can use AI to augment our problem-solving capabilities as humans, and help improve a lot of lives.

What other groups or activities are you involved with at Stanford?

I’m on the student board for United Students for Veterans’ Health, which is a nonprofit dedicated to improving the quality of life for military veterans in the VA system. These brave people made incredible sacrifices to protect us, and I’m honored to have the opportunity to spend time with them and hear about their life stories. Outside of USVH, I’m also involved with the Office of Community Standards and Organization Conduct Board, and play with Stanford Badminton.

Humans of Symsys: Antonio Aguilar

“This is becoming the theme of my blog post, but I’d advise students be intentional about who they are modeling, about developing relationships with the right models and mentors, and about asking oneself when it is best to go in a wholly new direction.”

Antonio graduated in 2018 and is a current Advising Fellow for the major, so he helps prospective and current SymSys students navigate their journey at Stanford. His office hours for Fall 2018 are Tuesday 9:30am-1:20pm in 460-040A.

What drew you to SymSys?

When I was growing up, I wanted to be a civil engineer like my grandfather. He built roads and bridges for the Costa Rican government for years. However, I got really into philosophy in high school. I started a weekly reading group which introduced me to the joy of discussing ideas between friends. I wondered then if I really should be pursuing a day job that was so separate from the intellectual life I wanted.

Before coming to Admit Weekend, I came across SLE and SymSys. I thought this Stanford combination sounded like the ideal liberal arts education for the 21st century, the perfect blend between technical disciplines and the human questions that really interested me. I looked through the SymSys core and realized that I would take all of those classes even if I didn’t need to fulfill the requirements of a major. SLE even knocked down Philosophical Foundations 1!

As my Admit Weekend came to a close, we ProFros got to hear from Srinija Srinivasan, one of Yahoo’s first employees and a SymSys alum. I remember being so impressed by her eloquence and the way she spoke about her own education that I practically locked down my decision right then and there.

Why did you pick your current concentration? Why did you decide to become an AF?

Many times I have made significant decision in my life, it has been because of role models. That is, I’ve seen something in someone else that has made me say “I want to be a bit like that”—Ms Srinivasan, for example.

So the answers to these two questions are actually the same. I attended the SymSys Student-Alumni-Faculty reunion during homecoming of my freshman year. (I highly recommend these, they are really fun!) There I talked to a junior called Cristian, who was an AF in the Artificial Intelligence concentration. The long conversation I had with him, sophomore Gerardo and alum Aman dispelled any doubts I might have had about SymSys and made me look further into AI for myself.

But just as important as that intellectual component was looking at someone who was joyful, Latin American, involved, excited about his work, and setting my nav roughly in that direction. I declared AI with Cristian, figuring I could change my direction later. It has taken more time, more models (people like Jure Leskovec in the CS department) and a 13th quarter to really commit to that path, but I am really happy for it. Getting involved in the SymSys community and working for it as an AF has been a phenomenal part of the process which was kickstarted that afternoon.

What’s your favorite SymSys-related class that you’ve taken?

It’s hard to choose! The list of non-SymSys-related classes I’ve taken is quite a bit smaller. The answer to that question is likely Phil 81: Philosophy and Literature. I think that class brings up essential questions about what makes a well lived life in relationship to art. In the realm of SymSys, I’ve yet to fully replicate the levels of flow and deep work I experienced while coding for CS106A or working through problem sets for Stats 116. The AI classes that I’m taking this quarter are giving me that feeling again, which makes me enormously excited.

What's one piece of advice you'd like to offer to younger students?

I came to Stanford as an international student from Costa Rica. Especially during my first years here, I looked at others around me to model how to choose my classes, much to work on them, how to spend my spare time. I see now that I didn’t always make the right choices for my particular circumstances and qualities and aspirations. This is becoming the theme of my blog post, but I’d advise students be intentional about who they are modeling, about developing relationships with the right models and mentors, and about asking oneself when it is best to go in a wholly new direction.

What (loosely) SymSys-related topic are you excited about right now?

Other than the AI stuff I’m studying and working on, I’m really excited that more people are becoming aware that technology often does not help us live the kind of lives that we truly want to live. I’m interested in better business models and design frameworks that help realign our technology with our humanity. I’m also invested in minimalism as a practical philosophy that seeks to bring our focus to what is truly important to us and brings us value.

What other groups or activities are you involved with at Stanford?

I was very involved in the Sigma Phi Epsilon community, especially with the project to house a part of our members in a multi-organization, multi-gender space. I have also been an active member of the Catholic Community at Stanford. Things change as a commuter, but I’m particularly excited to help kickstart the Thomistic Institute, a new offshoot of the Catholic Community that focuses on our intellectual tradition and the philosophy of Thomas Aquinas.

Humans of SymSys: Allen Nie

Allen Nie (MS '17) works on NLP and AI research with Professor James Zou in the Biomedical Data Science department at the Stanford School of Medicine. 

What drew you to the SymSys master's program? Why did you pick SymSys as opposed to other programs or related fields?

I've always been very interested in interdisciplinary fields. SymSys (unlike other cognitive science programs) has a strong focus on computation. It studies artificial intelligence and its surrounding fields without the need to focus on a particular approach (such as a cognitive science approach, symbolic/logic approach, statistical approach, etc.). It allows the students to understand the merits of all approaches and delve into the one they can believe in. In the age of artificial intelligence, I would like to see Symbolic Systems grow and become an actual department or institution with full-time researchers studying the mechanisms and fundamentals of intelligent systems and their behaviors.

What’s your favorite SymSys-related class that you’ve taken?

LING 130A with Chris Potts and EE 376A with David Tse. LING 130A focuses on understanding language using tools of logic and the later one focused on probabilistic systems and information theory.

Both classes teach a tool that allows students to understand aspects of the world. LING 130A teaches logic and EE 376A teaches probabilistic operators such as entropy and mutual information. Without getting into a philosophical abyss that is epistemology, the human scientific exploration focuses on featurizing the world, using models (sometimes our own brain) to find meaningful signals, discretizing the finding into digestible patterns or rules or “discoveries”, and call it knowledge. In order to study a field of interest, an analytical tool must be acquired. Both classes provide tools that allow us to model our hypotheses of the world and verify them.

What is one piece of advice you'd like to offer to younger students?

Find a field that you'd like to study. It can be music, language, cognition, nervous systems, molecules, networks of humans, and think about the best tool you can find to study them.

What's something cool you recently worked on? 

Universal sentence encoding, building interpretable NLP models, and ML safety where the model has the chance to abstain/reject when it is not confident about its decision. 

What underlying questions and issues do you hope to tackle/learn more about?

I've been working on universal sentence representation and I'm applying these representations to low-resource domains (a problem with very limited training data) such as analyzing clinical text. My other area of interest is to make NLP models more interpretable. In the past decades, we haven't had models that can truly match human-level performance on language tasks. But now we do. Is there hidden knowledge that we can learn from these high-performing predictive models? This is what statisticians have been doing in the past -- using a predictive model to conduct feature selection, and a clinician will jump out and spit out “definitive proof" about wine drinking and living long age. With the seriousness of correlation and causation aside, which may very well be part of a larger philosophical debate, are we able to discover valuable patterns on language from a high-performing predictive model with minimal inductive biases? Would those patterns contradict current linguistic theories? Sam Bowman (a Linguistics professor at NYU) has published some amazing work on this topic last summer.

As a diverse program with a lot of flexibility, many students struggle to find continuity across their coursework. (How) do you address this?

I don't think SymSys program provides continuity. Instead, each student needs to define the continuity from their own interests. If you are interested in language, you can follow the path of studying language and acquiring various tools, but you need to find the common threads among the classes you take. A professor in a statistics class won't tell you that Andrey Markov was thinking about fitting the vowels and consonants in Pushkin's poetry after he developed Markov Chain, or Charles Spearman, who invented Spearman correlation, is a highly revered Psychologist and regarded his own statistics work as secondary to his quests on Psychology (human intelligence). In order to find continuity in your work, you should first find out your personal interest and construct your curriculum and research around it.

If you could go back in time and be a Stanford student again, what would you have done differently and why?

There have been classes that I'm very interested in but was not able to take, such as MKTG 355: Designing for Happiness or CS 181: Computers, Ethics, and Public Policy — especially considering the Uber incident, who is to blame and to punish when a software that seemingly possesses agency kills someone? And since machine learning is marching into healthcare, is a line such as “our system on average only makes mistakes and kills 1000 people (less than 5000 people in a human-based setting) in your hospital each year” defendable in the public scrutiny?  I imagine I might have gotten an answer if I took CS 181 :)