Recommender systems are the algorithms that determine what content we read, which products we buy, which movies we watch. However, most of them are based on similarity and lock us into “filter bubble” where we see only what we expect. In this article we discuss how to bring in more serendipity into the algorithms.
Excel was a revolutionary product for the time, but in the 21st century columns and rows are not sufficient anymore. What matters today are multidimensional relations between data, which combine a better view of the bigger picture with an attention to detail: networks and graphs.
A specification for multi-purpose graph database structure, with examples and an online web / mobile implementation using Neo4J graph database and Node.Js.
Anything can be an interface, as long as it is in between and as everything is in between, everything is an interface. Interfaces create affordances or possibilities for action. They affect our perception and cognition. How can the interfaces with such transformative potential be designed? Moreover, if interface create a set of actions, what if we go beyond the notion of interfaces and think in terms of the protocols? Not the kind of protocols that say to do this and not to do that. The kind of protocols that transcend reality, that lie in the realm of practices, such as meditation, artistic practices, version-control systems, BitCoin, chaotic itinerancy, and other polysingular approaches.
Currently we have many tools that help us retrieve and store knowledge. However, there is a lack of tools that help us connect and make sense of all the disjointed bits of knowledge. Here we propose a certain cognitive framework, which could be used to tackle the problem of fragmentation, put information into the context, making it easier to not only find the right answers (something that software can already do well), but also discover the right questions.
