David
Graus

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.

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:

• 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.

I was invited by UvA’s Information, Communication and the Data Society (ICDS) to participate in a panel at the Conference on Privacy and Data Protection, which was focused on AI.

The recording of the panel is now online, watch me telling a room full of (highly) privacy-aware (and cookie-averse) people that Cambridge Analytica nudging people to “politically activate them” with tailored information can be a “democratic good” 😅.

See the recording below:

For more information, see CPDP’s page of the panel.

For the Media Perspectives (“platform for media & innovation”) newsletter I wrote a column on the future of media & AI: Algoritmewijsheid. Read it below (in Dutch):

Enjoyed giving a lecture at the SIKS Course “Advances in Information Retrieval” at the Mitland Hotel in Utrecht. I also pitched DIR 2019 😅 (as evidenced by the picture above from Arjen). See my slidedeck below!

[slideshare id=180305564&doc=siks-advances-in-irbias-in-recommendations-191009080815]

This talk is loosely based on (part of) the talk I gave at the ACM RecSys Summerschool, but I added a few slides on dealing with implicit feedback (= clicks), and popularity bias.

With Daan Odijk I gave a lecture + hands-on workshop at the ACM Summer School on Recommender Systems in Gothenburg, Sweden on RecSys in the Media Industry: Relevance, Recency, Popularity, and Diversity.

For it, we had a long (90+ min) lecture combining insights, experiences, and projects from our work at RTL and Blendle (Daan), and FD Mediagroep (me).

In addition, we did a small hands-on workshop, implementing a content-based re-ranker for WikiNews.

See our slides and notebooks here: https://github.com/graus/recsys_summer_school/

See a tweet by @alansaid, here:

Day three of the @ACMRecSys summer school is on. @dodijk and @dvdgrs talking about #recsys in media. #recsys2019 pic.twitter.com/NRV1H7A0So

— Alan Said (@alansaid) September 11, 2019

Finally, see my slidedeck here:

[slideshare id=171560450&doc=2-graus-fd-mediagroep-recsys-summer-school-190913133918]

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

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)!

🥇 Content

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

“Finally 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)

🥈 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 😅, these are the awards we won with SMART Radio:

🏆 Marconi Online Award
🏆 AMMA Award (with SMART Journalism)
🏆 DIA2019 Award, category: Content
🥈 DIA2019 Award, category: Disruptor

In the context of the upcoming workshop “AI and Journalism” Bennie Mols interviewed me for the website of the European Science-Media Hub, on how we apply AI in journalism for Het FD and BNR. Read the piece here:

• AI could boost journalism with smart tools

Or read the more in-depth background interview here:

• A scientist’s opinion: Interview with David Graus about AI-tools for journalists

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 🤖. 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 🤓