Let’s build networks of science-friends!

Conference fatigue. I’m realizing that conferences are not really working for me at the moment. In spite of heroic efforts from conference organizers, super-star speakers, etc. When experienced through the screen in my spare bedroom, it’s all turning into a uniform, gray blur. And I generally don’t enjoy giving talks into a matrix of empty, muted windows.

There is a deeper problem: But who cares about privileged old Sune? It has been argued – and I strongly agree – that the key victims of conferences moving online are the young researchers. The PhD and PostDoc level.

By not attending in person, the young researchers are robbed of ways of building networks of “science friends”. Other young researchers in other labs and countries that they’ll need as their careers progress. (To this day, many of my best science-friends are people I met at NetSci conferences during the aughts).

And this is also a loss for Science more generally. What will our fields look like as these networks are no longer allowed to grow.

Restoring network-building for young researchers in the age of COVID-19 lockdowns. So maybe we should work to set up structure to run talks/conferences in a way that favors the young researchers. Let’s build networks of science-friends!

I’ve thought about this for a while. And gotten good ideas from the actual young researchers in my group. Here’s the general idea:

  • For increased engagement, we need to shift the focus to discussions rather than talks. So we need small groups.
  • I would still make it about presenting papers. But we need to do it in a way where people don’t just surf the internet during talks. For that we need small groups. And lots of discussion afterwards.
  • It’s about letting friendships grow. For that we need small groups. And lots of discussion.
  • Conferences are also about “being seen” by senior researchers in the field, so maybe we should put one senior person in each little group.

The nitty gritty: To make this concrete, I would

  1. Start by collecting a long list of interested young researchers.
  2. Get research keywords from everyone.
  3. Group people into small groups of 8, with groups based on shared interests, but forming links across labs, institutions, and countries
  4. Groups could meet weekly for 1-2 hours.
  5. At each meeting there would be 2 short talks, then discussion.

After 4 weeks, the group would be done. And we could do multiple runs to foster more connections. Build more possibilities for friendships. So everyone would have a crew to have dinner with once the real conferences resume. (How I miss those!).

Looping in the seniors: But wait a minute!!!

Another important aspect of conferences (from the perspective of young researchers) is to meet famous people from their field. To see those authors of great papers in person.

Thus, to supplement, we could also gather a list of senior researchers, who would volunteer to sit in on the talks, provide a bit of initial feedback and get the discussion started.

The oldies should rotate every week so that the young people would have a chance to hear from the max number of different senior people.

Improving: Let me know what you think! If people like this idea, maybe we should get something like this going? I’m especially interested to know.

  • Do you have any ideas for improvements of this format?
  • If you’re a young researcher, does this sound interesting to you?
  • If you’re a senior researcher would you be up for joining as a mentor?

(Thanks to the researchers in my group for feedback on an early version of this post)

Contact Tracing

Update May 23rd. Final update for this post. I’m happy to report that the Danish government – in part based on input from our advisory board – has decided to base the Danish contact tracing app on the DP-3T (as implemented by Google and Apple) framework. Details here https://www.sum.dk/Aktuelt/Nyheder/Coronavirus/2020/Maj/Politisk-aftale-om-frivillig-smittesporingsapp-for-covid-19.aspx

Update May 1st. Here’s another update. It also turns out the advisory board wasn’t officially announced until today. I hope the unintended embargo-breaking below won’t have any serious consequences. Here’s the official press release: https://digst.dk/nyheder/nyhedsarkiv/2020/maj/nyt-advisory-board-skal-raadgive-myndighederne-om-den-kommende-danske-smittestops-app/ see also here.

Update: April 27th, 2020. Well, it turns out I was wrong about Reason 1 below. And that I have lousy timing in writing blogposts with declarations in them.

Pretty much right after finishing the post below (original post was written on April 26th), I was invited to be a member of a newly established Advisory Board for the official Danish Contact Tracing App. The goal of the advisory board is to provide advice/feedback for the Agency for Digitization and the Ministry of Health regarding the App.

Since I care a lot about doing digital contact tracing in the right way (with respect to privacy & epidemiological relevance), I accepted the offer to join. Clearly that changes some aspects of what I’ve written below.

Over the past few weeks, I’ve commented on contact tracing in the Danish press. But starting today, I am going to stop chiming in on this topic.

Reason 1 is that I don’t feel like I’m making much of an impact in terms of where the official app is going (which was my main reason to start speaking about this in the first place). Reason 2 is that the news on this topic are moving fast and staying updated is exhausting. Reason 3 is that engagement in this topic is drawing me away from the deep & long-term work that I want to be focusing on. And finally, Reason 4 is that contact tracing connects deeply to projects I did years ago, so it feels a bit removed from my current core scientific work of network science and mobility modeling. Thus, I’ll be doing my best to pipe down from now on.

If you want updated info about contact tracing in Denmark, I recommend checking out Henrik Moltke’s twitter feed. For the overall (epidemiology and privacy) view on contact tracing, I personally follow Marcel Salathe at EPFL.

Here are my two key points on contact tracing:

  • A decentralized protocol (such as DP-3T) is the best way to do things.
  • Contact tracing is not necessarily the best way to counteract spread of COVID-19. (For example: Re-focusing all the resources we’re currently using on building apps towards getting everyone to wear masks, might very well be a better strategy.)

Below, I provide some context on each of these points, referring back to what I’ve already said in the press.

Privacy preserving contact tracing

In the beginning of the epidemic, I supported (and made a small contribution to) the PEPP-PT initiative, which was then more of an umbrella organization, also including DP-3T. See this press release from DTU. At some point there was an internal argument (which I know nothing about), resulting in a split into a centralized (still privacy preserving) paradigm embodied by PEPP-PT and a decentralized paradigm represented by DP-3T and also incorporated in the Apple/Google solution to contact tracing.

I have provided continuous updates on my attitude towards this issue via Twitter. See below

I’m placing an image instead of embedding the tweets because I can’t figure out how to display the thread. Link here https://twitter.com/suneman/status/1251101267956502531
Again, I’m putting an image here instead of an imbed – since I can’t figure out how to embed the thread. But you can find it here: https://twitter.com/karmel80/status/1253256268283891713

Here’s a link to a nice report on issues with centralized solution used in PEPP-PT. This report, written by members of the DP-3Tproject, came out shortly after PEPP-PT finally provided open source code.

Is contact tracing using cellphones even a good idea?

An important point to make when it comes to contact tracing is that it might not be a very good solution. Many of the issues are discussed in this Twitter thread by UNICEF Machine Learning Lead & Principal Researcher Vedran Sekara.

I’ve also talked about these issues in the Danish Press.

Achievement Unlocked!

Starting march 1st this year, I was promoted to full professor at DTU! Pretty exciting. And if you want proof, take a look at my updated profile page (conveniently screen-shot and marked up below).

Hmm. And I should figure out how to get a new photo – this one is from 2009.

Arek Stopczynski Visit

We’re lucky to have Lab Alum Arkadiusz ‘Arek’ Stopczynski visiting the lab on January 9th and 10th. On the 9th, he’s busy being examiner at a PhD Defense, but on his second day in Denmark, he’s going to give a talk to tell us about what he’s been up to since starting at Google af couple of years ago.

In addition to working with us at DTU, Arek has also been a postdoc in Sandy Pentland’s lab at MIT’s MediaLab, he was an integral part of building the world’s first mobile brain scanner, and he’s given a great TEDx talk. And did I mention that we’ve just put out yet another paper together.

Below, you can find the talk details:

  • Date: January 10th, 2019
  • Time: 13:30-14:30
  • Place: DTU Building 321, in the first floor lab space
  • Title:  Data Science: Thinking Industry
  • Abstract: The practice of Data Science involves employing different methodologies, techniques, and tools, both deconstructive and constructive. In this talk we will discuss some fundamental differences in how Academia and Industry (exemplified by large tech companies) approach teaching and applying Data Science. These differences have important implications for how we teach students and conduct research.

I hope you can make it. Arek will stick around afterwards if you’d like to chat and hang out.

Complex Networks in Cambridge

I had an absolutely wonderful time at the Complex Networks 2018 conference last week in Cambridge, UK. I learned a lot and got caught up a bit with all the amazing work that’s going within complex network analysis and see some of the great new young researchers in the field.

At the community detection sessions, I also saw several talks that drew on our work on Link Clustering, expanding and building on those ideas. Now don’t get me wrong: That work is well cited, so I know people have been reading it. But my sense is that most of the citations are of the type “This is also something one could do” or from people applying the algorithm. Those are both great (and a much better fate than what befalls most of my papers), but it is still extra exciting to see people adopting, refining, and developing the ideas – using them for their own work with community detection methods!

Another exciting development was to see how lots of people are starting to apply machine learning (including embeddings, etc) to networks.

Finally, I also got to give my own keynote about our recent paper on the Chaperone Effect in Scientific Publishing. It was a brand new talk (since the paper just came out 2 days prior), but judging from the Twitter reaction, people liked it 🙂

TEDx Aarhus

Pantelis Pipergias Analytis visit

On October 9th, we are lucky to have Pantelis Pipergias Analytis visiting the group. Pantelis recently moved as an assistant professor at the Danish Institute of Advanced Studies (D-IAS) at the  University of Southern Denmark.

Before moving to Denmark, he spent the past two years as a postdoctoral researcher at the Computer and Information Science department at Cornell University. Pantelis got his PhD from the Max Planck Institute for Human Development in Berlin.

Pantelis will give a talk based on his recent Nature Human Behavior paper Social learning strategies for matters of taste

  • Date: October 9th
  • Time: 13:30
  • Place: Technical University of Denmark, Building 321, Room 134

TitleSocial learning strategies for matters of taste

Abstract: Most choices people make are about ‘matters of taste’, on which there is no universal, objective truth. Nevertheless, people can learn from the experiences of individuals with similar tastes who have already evaluated the available options—a poten- tial harnessed by recommender systems. We mapped recommender system algorithms to models of human judgement and decision-making about ‘matters of fact’ and recast the latter as social learning strategies for matters of taste. Using computer simulations on a large-scale, empirical dataset, we studied how people could leverage the experiences of others to make better decisions. Our simulations showed that experienced individuals can benefit from relying mostly on the opinions of seemingly similar people; by contrast, inexperienced individuals cannot reliably estimate similarity and are better off picking the main- stream option despite differences in taste. Crucially, the level of experience beyond which people should switch to similarity- heavy strategies varies substantially across individuals and depends on how mainstream (or alternative) an individual’s tastes are and the level of dispersion in taste similarity with the other people in the group.

Piotr Sapieżyński on Fairness in ranking

Our old friend Piotr, current postdoc at Northeastern, and graduate from the group is visiting from his new home beyond the Atlantic. This coming Thursday, Piotr will give a short about his most recent work. Details below.

  • Time: Thursday, Sept 6th. 11AM
  • Location: Technical University of Denmark.B321, lab-space
  • Title: Fairness in ranking

Abstract: Ranked lists of persons and items are a core part of the user experience in many online services, such as search, social media feeds, hiring, and dating sites. Studies have shown disparate amount of attention received by high rank results, potentially leading to loss of opportunity and access to resources among the lower ranked items. In this short talk I will give an overview of the work on individual and group fairness in ranked lists and focus on our work in progress: a novel metric for investigating group unfairness in ranked lists. Our approach relies on estimating the amount of attention given to members of a protected group and comparing it to that group’s representation in a defined population. It offers two major developments compared to the state of the art. First, rather than assuming a logarithmic loss in importance as a function of the rank, we allow for attention distributions that are specific to the audited service and the habits of its users. For example, more items are consumed in a single viewing of a social media feed than as a result of a single query in a web search engine. Second, we allow non-binary protected attributes (gender, race, etc.), both to better reflect the way individuals identify, but also to enable measurements on aggregates of multiple search runs, rather than separately for each result list.We investigate the properties of the metric and compare them to the behavior of other established approaches using synthetic ranked lists. Finally, we showcase the metric through a simulated audit of a number of hiring and dating services.

Bernardo Huberman and AI for the Network

Later this month we will have legendary researcher Bernardo Huberman visiting. And we’re lucky enough to have him giving a talk on one of the most exciting new developments in Network Science: Applying AI to networking problems.

Bernardo has been a central player throughout the rise of network theory (and mentor for field notables, such as Lada Adamic and Jure Leskovec), but that’s just a fraction of what he’s accomplished. If you care about anything related to information sciences, this is a talk you cannot miss.

Bernardo is a Fellow and vice president of the Core Innovation Team at CableLabs. He is also a Consulting Professor in the Department of Applied Physics and the Symbolic System Program at Stanford University. Previously he was Senior Fellow and Senior Vice President at Hewlett Packard Enterprise Company, and Director of the Mechanisms and Design Lab at Hewlett Packard Labs.

  • Date: August 29, 2017.
  • Time: 14:00
  • Location: Technical University of Denmark, Building 321, 1st floor: Room 134

Title: Artificial Intelligence and the Network

Abstract: Artificial Intelligence is the attempt to make computers emulate human cognition and thought processes. It has existed for a long time and has sprouted a number of subfields, from semantic networks and common sense reasoning to robotics, logic programming and machine learning. In spite of the glacial rate of progress in AI,  one subfield, machine learning, has recently taken off like wildfire. What powers this incredible growth is the availability of fast processors that have made possible computations than seemed hard to achieve a few years ago. As a result, we now have powerful systems that can easily recognize myriad images and spoken languages. This talk will describe some of the great successes of machine learning, their limitations, and their application to networking problems which pervade modern communications. I will also present a form of artificial intelligence that is distributed in nature and that mimics the ability of groups of people and social insects to solve extremely hard problems.

Max Schich Talk

We’re lucky to have Max Schich visiting DTU tomorrow. Max is an associate professor for arts and technology at The University of Texas at Dallas and a founding member of the Edith O’Donnell Institute of Art History. His work converges hermeneutics, information visualization, computer science, and physics to understand art, history, and culture. Schich is the first author of “A Network Framework of Cultural History” (Science magazine, 2014) and a lead co-author of the animation “Charting Culture” (Nature video, 2014). He is an editorial advisor at Leonardo Journal, an editorial board member at Palgrave Communications (NPG), and the Journal for Digital Art History. He publishes in multiple disciplines and speaks to translate his ideas to diverse audiences across academia and industry. His work received global press coverage in 28 languages.

Details

  • Time: April 17th, 14:00
  • Location: DTU, Building 321, first floor lab space
  • Title: Towards a Morphology of Durations

ABSTRACT: History has no periodic table of elements and no theory of temporal structure, as George Kubler pointed out in 1962, yet, as he also points out, things occupy time in a bounded number of ways. The obvious question still is: Can we capture the shape of time? – Tackling this challenge, this talk looks at historical time systematically, dealing with more or less exponential growth, the archaeological paradox, global and meso-level patterns, cycles, periodicity, condensation, and a bouquet of oddities.

Here’s a cool video about some of Max’s recent work