RT @oneDylanAshley: Have you ever wondered what happens when you reduce a story to a low-dimensional latent representation? Well my new IEE….

We built a tool where you can upload your playlist or album and sort it according to the different narrative prototypes. 7/n

This statistical evidence led us to conduct a user study, where we found that playlists sorted by these narrative prototypes were strongly preferred over shuffled playlists. While most of our empirical results focus on music, we also show similar results for movie stills. 6/n

This allows us to find simple, prototypical narrative arcs that are present for example in music albums. There is strong statistical evidence that music albums tend to follow curves similar to the ones we found. 5/n

In other words, it is the feature that is most predictive about what place in a story an item can take. We find that often, a one-dimensional feature is sufficient to capture the narrative information. 4/n

Using a machine learning approach, we learn to distill each item's narrative essence. Narrative essence is the feature that maximizes mutual information between the unordered collection of these features and the collection’s order. 3/n

When multiple items such as words, pictures, songs or emojis are put together to form a whole, they convey a story. If the story is the structure that connects all the items together, the narrative information of each item is its mutual information with the whole story. 2/n

Excited to share new work done with @oneDylanAshley, Zachary Friggstad and @SchmidhuberAI on stories, music, movies and information theory, published in IEEE TPAMI📚🎶 🎬🤖 Below a quick summary and a web tool you can use to sort your playlists:) 1/n

RT @FaccioAI: Had a great time presenting at #ICML2024 alongside @idivinci & @LouisKirschAI. But the true highlight was @SchmidhuberAI hims….

RT @damianborth: Pleasant surprise of @icmlconf: There is a growing Weight Space Learning Community out there - it was great to meet you al….

RT @Sam_kuyp: To celebrate @DavidDeutschOxf’s 70th birthday, colleagues and friends of David have contributed text and videos in which they….

RT @giffmana: lol, draft of full GPT-4 paper with architecture and data details is already leaked on torrent😂. The vision component in the….

RT @hardmaru: New paper from IDSIA motivated by building an artificial scientist with World Models!. A key idea is to get controller C to g….

RT @SchmidhuberAI: We address the two important things in science: (A) Finding answers to given questions, and (B) Coming up with good ques….

RT @martinmbauer: It pains me to see eminent colleagues like Neil deGrasse Tyson propagate elementary mistakes. What would really be the….

RT @SchmidhuberAI: Machine learning is the science of credit assignment. My new survey (also under arXiv:2212.11279) credits the pioneers o….

RT @QuantaMagazine: The 2023 Breakthrough Prize in Fundamental Physics was just awarded to four pioneers of quantum information science: Ch….

RT @FaccioAI: @SchmidhuberAI @idivinci @LouisKirschAI @darl_icml It builds on recent work where we learn a small set of abstract states tha….

RT @JanelleCShane: Stunning transcript proving that GPT-3 may be secretly a squirrel. GPT-3 wrote the text in green, completly unedited! ht….

RT @annalapwood: I think I have just discovered the best composer name of all time.