Dutch Interactive Awards (2x) for BNR SMART Radio!

πŸ“… June 28, 2019 β€’ πŸ• 10:17 β€’ 🏷 Blog β€’ πŸ‘ 52

Yesterday BNR SMART Radio won two Dutch Interactive Awards (DIA 2019) Awards! We were nominated together with Elastique, the folks who designed the UX/interface of our BNR SMART Radio app which you can download here (or in your app stores, both Android/iOS)!

πŸ₯‡ Content

We won Gold in the category “Content”. Why, you may ask?

β€œFinally a reason to listen to Radio in general, and BNR in particular. Because advertising is absent and topics are tailored to the personal taste of the listener, who can search in a targeted way and express appreciation for a broadcast through a thumbsup, they reach new target audiences online.”

jury report (source: emerce)

πŸ₯ˆ Disruptor

Next to this, we also won the Silver DIA2019 Award in the category “Disruptor”!

“In a world where radio has never deviated from the linear model, the jury finds this very disruptive. SMART Radio provides a personal radio experience and possibly a new revenue model in the long term. The jury members find the potential of the underlying strategy even more impressive. What if BNR joins forces with FD Mediagroep and combines the work of both parties in this? The jury sees all kinds of opportunities to fit in new forms of advertising (or to omit them). The collaboration between the agency and the media company has resulted in something that has never been seen before.”

jury report (source: emerce)

Current standing

For who has trouble keeping track πŸ˜…, these are the awards we won with SMART Radio:

πŸ† Marconi Online Award
πŸ† AMMA Award (with SMART Journalism)
πŸ† DIA2019 Award, category: Content
πŸ₯ˆ DIA2019 Award, category: Disruptor

BNR SMART Radio wins the Marconi Online Award!

πŸ“… February 1, 2019 β€’ πŸ• 16:17 β€’ 🏷 Blog β€’ πŸ‘ 90

Really happy and proud of our amazing AI team for winning the (first) Marconi Online Award at the Gouden RadioRing Gala, for BNR SMART Radio!

SMART Radio is the first product that comes out of our FD Mediagroup AI team! SMART Radio is released in beta for iOS (anytime) and Android, give it a try by downloading it here: https://bnr.nl/smartradio.

Image: BNR

Read more about SMART Radio in our demo paper, presented at DIR 2018:


Improving automated segmentation of radio shows with audio embeddings published @ IEEE ICASSP2020

πŸ“… February 12, 2020 β€’ πŸ• 11:54 β€’ 🏷 Papers β€’ πŸ‘ 46

Oberon Berlage’s MSc. thesis: “Improving automated segmentation of radio shows with audio embeddings” which he wrote under my supervision during his internship at FD Mediagroep was awarded a 9/10, under condition that the work was publishable.

Turns out it was, as it was recently accepted at IEEE ICASSP2020 (the 45th International Conference on Acoustics, Speech, and Signal Processing) without any additional work/experiments (just a bit of reduction). But you already knew this… Oberon will be presenting this work in Barcelona, thanks to the generous support of UvA’s Information Studies program.

We now published a preprint, read it below:

  • [PDF] [DOI] O. Berlage, K. Lux, and D. Graus, “Improving automated segmentation of radio shows with audio embeddings,” in Icassp 2020 – 2020 ieee international conference on acoustics, speech and signal processing (icassp), 2020, pp. 751-755.
    [Bibtex]
    @inproceedings{berlage2020improving,
    author={O. {Berlage} and K. {Lux} and D. {Graus}},
    booktitle={ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
    title={Improving Automated Segmentation of Radio Shows with Audio Embeddings},
    year={2020},
    pages={751-755},
    doi={10.1109/ICASSP40776.2020.9054315},
    url={https://doi.org/10.1109/ICASSP40776.2020.9054315}
    }

His work revolved around improving BNR SMART Radio‘s text-based segmentation by incorporating audio signals in the form of audio embeddings. This turns out to improve over our text-based baseline by a whopping +32.3% F1-measure!

Even better: an audio-only approach, trained on a smallish openly available dataset, outperforms our text-only baseline by 9.4%. This means the segmentation method can be employed without need for audio transcription, which could be a money-saver.

“Improving automated segmentation of radio shows with audio embeddings”

πŸ“… July 5, 2019 β€’ πŸ• 15:41 β€’ 🏷 Blog and Research β€’ πŸ‘ 74

Update (28/1/2020): Oberon’s thesis was accepted and will be published at the IEEE 45th International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2020), to be held May 4-8 in Barcelona, Spain! The submission is co-authored with Klaus Lux and myself.

Oberon Berlage recently successfully defended his MSc. thesis (title above!) for the Data Science Master at University of Amsterdam, and graduated with a whopping 9!

He’s the first academic offspring of our AI Team @ FD Mediagroep, and worked on BNR SMART Radio‘s segmenter. Oberon improved our text-based segmenter by adding audio embeddings, improving the F1 score with +32%!

His thesis is now online, check it out at: http://scriptiesonline.uba.uva.nl/document/673254

Featured in BNR’s “Artificial Intelligence Podcast” on AI in Media Part I: Tech

πŸ“… November 27, 2018 β€’ πŸ• 10:22 β€’ 🏷 Media β€’ πŸ‘ 19

In the latest AI Podcast, I am featured, talking with Li’ao Wang on our SMART Radio and SMART Journalism projects. Stream the episode below!

“Media is smart, but needs to get smarter. That’s what we’re working on here at the FD Mediagroep. Come and see what we do with AI to make our creators and content better.”

Stream on BNR

Stream on Spotify

PodRecs: Workshop on Podcast Recommendations PC

πŸ“… April 4, 2020 β€’ πŸ• 12:14 β€’ 🏷 Research β€’ πŸ‘ 2

I was invited to join the program committee of (the first) PodRecs: Workshop on Podcast Recommendations (to be held at RecSys’20).

Since our work on BNR SMART Radio, I am really interested in the space of audio, recommender systems, and information retrieval. Curious to see the submissions!

See the PodRecs call for papers, and check out the website, by clicking the image below.

We won the AMMA Media-innovation award for our News Personalization projects at FD Mediagroep!

πŸ“… May 17, 2019 β€’ πŸ• 13:50 β€’ 🏷 Blog β€’ πŸ‘ 44

After winning the Marconi award for radio innovation, yesterday we picked up the AMMA Award for Media Innovation for our news personalization efforts at FD Mediagroep (both SMART Journalism for Het FD and SMART Radio for BNR Nieuwsradio)!

It’s really great to see that our current investment into AI and innovation seems to resonate with the outside world πŸ€–. And I am really happy to be with a company that sees this development as such an important direction that we are able to work with a big and talented team of data scientists, interns, engineers, and product folks πŸ€“

Press

πŸ“… September 22, 2017 β€’ πŸ• 13:14 β€’ 🏷 β€’ πŸ‘ 163

Below are a selection of press articles about me and/or my work. For a bit more context on some of the publications below, check out my posts in the media category.

Legend: 🌐 Online | πŸ“° Paper | πŸŽ™ Podcast | 🎦 Video

2020

  • 🌐 6 May | Villamedia: Het idee van de filterbubbel kan in de prullenbak
    Column that mentions our UMAP 2020 paper. link

2019

  • 🌐 28 November | Media Perspectives: Algoritmewijsheid
    Column by me, on the future of AI and media. link
  • 🌐 4 June | European Science-Media Hub: A scientist’s opinion: Interview with David Graus about AI-tools for journalists
    Interview on how we apply AI in Journalism (SMART Journalism & SMART Radio). link
  • 🌐 4 June | European Science-Media Hub: AI could boost journalism with smart tools
    Piece on AI in Journalism, in the context of an upcoming workshop. link
  • 🌐 10 May | De Nationale AI Cursus: Geef niet het algoritme de schuld
    Interview in response to my op-ed in Het Parool. link
  • πŸ“° Mar 23 | Het Parool: Antivaccinatieboeken
    Published letter in Het Parool on whether bol.com’s selling of anti-vaxx books is an algorithmic issue (it’s not). link
  • πŸ“° 8 Mar | De Volkskrant: Nog even en de robotjournalist schrijft ook voor Nederlandse media
    Interview about robo journalism in Dutch, with a special mention of our SMART Journalism project. link

2018

  • πŸŽ™ 26 Nov | BNR: The Case of: AI in Media Part I – Tech
    Interview for BNR’s AI Podcast where I talk about SMART Journalism. link
  • πŸ“° 12 Oct | De Volkskrant:Β Met deze eigen zoekmachine spit de politie schatten aan digitaal bewijs door
    Interview about the searchengine developed by the NFI, and its use in the high-profile Mocro maffia court cases, with my view on its role in the collection digital evidence. link
  • πŸ“° 9 Oct |Β IP – vakblad voor informatieprofessionals: Vensters op een nieuwe werkelijkheid.Β [image]
    Interviewed as a ‘young professional’ that show that traditional categories and thinking frames need to be readjusted, where I explain how my background and cross between ‘hard’ and ‘soft’ science helps in my current work in AI and media.
  • 🌐 10 Oct | Villamedia: De journalist en de nerd
    Much needed and insightful nuance by Nick Kivits in an op-ed in response to our three-way interview on Villamedia (below). Nick advocates for IT folks and journalists to meet for coffee and collaborate, instead of having mutual suspicion or distrust. link
  • 🎦 6 Oct | RTL XL: How it’s done
    Television show about intelligence and business information. In this video me and Company.info’s CTO Henk Pijper discuss our work in AI and news. YouTube | blog
  • 🌐 5 Oct | Villamedia:Β Algoritmes: Moeten journalisten vrezen voor hun toekomst?
    Three-way interview with myself, Anne Schuth (De Persgroep), and Daan Odijk (RTL Nederland) about the rise of AI in media. link
  • πŸ“° 2 October | NRC:Β Waar blijft de Nederlandse robotjournalist?
    Article on ‘robo-journalism’ (again), where I mention our SMART Journalism project. link
  • πŸ“° 10 July | Communications of the ACM (CACM): Dutch Police Fight Crime by Cracking PGP Phones
    Article about my work for Inez Weski on the Noffel case. link
  • 🌐 1 May |Β Stimuleringsfonds voor de Journalistiek:Β Waar blijft de robotjournalistiek in Nederland?
    Article about ‘robo-journalism’, where I say something about the SMART Journalism project we are doing at FDMG, which involves personalization and summarization of newspaper articles. link
  • 🌐 18 Apr | Nemo Kennislink: Het sleepnet van Justitie
    Article about the research I did into the digital forensic methods of the NFI/Dutch Public Prosecutors in the high-profile “Noffel” case. link

2017

  • 🎦 12 Nov | Denktank: Baas over brein
    My debut on national TV ;-). Children’s TV show, on future topics, in this episode: how companies influence us online through persuasion/behavior analysis. [video | snippet (01:45min)]
  • πŸŽ™ 10 Nov | New Business Radio: Let’s talk business
    Podcast on doing data-driven sales and marketing with Company.info. [link | podcast]
  • πŸ“° 3 Jul | FD: Wanneer je gegevens geld waard zijn [link]
    Article on the value of personal data in which I’m asked a few questions on how user interactions are valuable to tech companies.
  • 🌐 17 Jun | NOS:Β David onderzocht de sporen die wij onbewust online achterlaten [link]
    Article on my research/dissertation: Entities of Interest — Discovery in Digital Traces.
  • πŸ“° 15 Jun | Folia:Β Deze promovendus weet alles over jouw online voetafdruk [link]
    Article on my research/dissertation: Entities of Interest — Discovery in Digital Traces.

2016

  • πŸ“° 17 Jun |Β NRC: Wij zijn racisten, daarom Google ook [link]
    Op-ed on why it’s a good thing that algorithms aren’t neutral.
  • πŸ“° 10 Apr | Tekstblad: Waarheidsvinding in 11,5 miljoen documenten
    Article on my research: Discovery in Digital Traces.

2014

  • πŸ“° 17 Apr | Emerce: Nederlands algoritme reikt nieuwe Wikipedia onderwerpen aan [link]
    Article on our ECIR 2014 paper:Β Generating Pseudo-ground Truth for Predicting New Concepts in Social Streams.
  • 🌐 17 Apr | Business Insider:Β Algoritme gebruikt Twitter om nieuwe Wikipedia-artikelen te voorspellen [link]
    Article on our ECIR 2014 paper:Β Generating Pseudo-ground Truth for Predicting New Concepts in Social Streams.
  • 🌐 17 Apr | UvA:Β New method supplements Wikipedia with Twitter topics [link]
    Article on our ECIR 2014 paper:Β Generating Pseudo-ground Truth for Predicting New Concepts in Social Streams.
  • πŸ“° 16 Apr | DeMorgen.be: Nieuw algoritme vult Wikipedia aan op basis van tweets [link]
    Article on our ECIR 2014 paper:Β Generating Pseudo-ground Truth for Predicting New Concepts in Social Streams.
  • 🌐16 Apr | Tweakers.net: Nieuw algoritme vult Wikipedia aan op basis van Twitterberichten [link]
    Article on our ECIR 2014 paper:Β Generating Pseudo-ground Truth for Predicting New Concepts in Social Streams.

About

πŸ“… June 14, 2010 β€’ πŸ• 12:28 β€’ 🏷 β€’ πŸ‘ 2,238
foto: BeeldstudioKB

I am an AI practitioner. I lead the data science chapter at Randstad, the global leader in the HR services industry. With 13 data scientists spread out over different product teams we work on projects involving recommender systems for job matching, NLP and information extraction. From 2018 to 2019 I worked as lead data scientist at the FD Mediagroep, where I led a team of five data scientists in building a custom learning to rank-powered recommender system for het Financieele Dagblad (the Dutch equivalent of the Financial Times), and the award-winning SMART Radio for BNR.

I have an academic background. I acquired my PhD degree in Information Retrieval from the University of Amsterdam (UvA) in 2017, where I worked under supervision of prof. dr. Maarten de Rijke at the Information and Language Processing Systems Group (ILPS).

My research revolved around semantic search and computational methods for automated understanding of large-scale textual digital traces. I defended my thesis “Entities of Interest β€” Discovery in Digital Traces” in June 2017. In the summer of 2015 I did an internship at Microsoft Research in Redmond, under supervision of Paul Bennett, Ryen White, and Eric Horvitz, where I worked on analyzing Microsoft’s Cortana user-interaction logs, which resulted in a best paper award at UMAP 2016, and the filing of a patent. In the winter of 2014 I visited prof dr. Doug Oard‘s E-Discovery Lab at the University of Maryland, College Park.

Algorithms aren’t evil

In my opinion the voice of computer scientists is missing in the highly polarized (public) debate around AI, which is largely dominated by non-technical people. For that reason, I think we wrongly fear filter bubbles, think personalized advertising will usher in the the end of democracy, believe tracking cookies are our worst enemies, and feel that companies should not be allowed to collect any of our data at all. I try to bring another perspective, and in that context I try to speak and write to get my point across to a broader audience. I have written and spoken in mainstream media for the general publicβ€”which I try to continue to do. Some things I am particularly proud of:

  1. An article on why it’s a good thing algorithms aren’t neutral, written with Maarten de Rijke, published in NRC Handelsblad and nrc.next.
  2. An appearance on TV in “Denktank,” a program for youngsters, where I nuance the worries around algorithms and algorithmic personalization.
  3. A layman’s talk on how and where algorithms affect us in daily life for a debate night themed “The power of algorithms” at De Balie in Amsterdam.
  4. A keynote at the VOGIN-IP Lezing titled “The filter bubble doesn’t exist,” where I explain algorithmic personalization and summarize a few academic works that suggest the filter bubble is not an algorithmic problem.
  5. Consulting lawyer Inez Weski in making sense of the digital forensics process that played a role in a large-scale lawsuit involving the acquisition of a huge database of encrypted communication data (Ennetcom).

Media Studies

I have a background in Humanities, and obtained my Media and Culture (media studies) Bachelor of Arts at the University of Amsterdam in 2008. After a brief period of work and travel, in 2009 I started my MSc Media Technology at the Leiden Institute of Advanced Computer Sciences, at Leiden Universiteit. In 2012, I graduated with a project involving semantic annotation and network visualization.

Professional experience

I used to work in the media. I’ve worked for Dutch broadcasting company NTR (formerly Teleac, former-formerly Teleac/NOT) for several years as an editor for radio and online science programs. Here I gained experience in a wide array of editorial tasks such as writing (news) items, preparing interviews, but also with several aspects of audiovisual media production, like (video) editing, animating, and studio recording of a daily science news program.

Science communication

During my PhD I leveraged this experience by writing short articles about our group’s publications for the general public, with the communications department of the University of Amsterdam. I co-authored and published several articles, some of which gained (inter)national coverage on tech sites, newspapers and magazines. See for example:

  1. New method supplements Wikipedia with Twitter topics (also appeared on Tweakers.net, DeMorgen.be, Emerce, z24.nl)
  2. Improved predictions of queries by search engines
  3. New method helps Google Translate translate better (also appeared in NewScientist)

What else?

I further possess basic skills in graphic- and web-design, having built several websites for several people, creating thesis covers for friends, and logos for projects. I like hacking away at my car. I like photography, which you can witness on Instagram (@instagraus). I also like yoga and running. And like the rest of the world I like to travel, read books, watch films, play videogames, and listen to music.