Recently developed information communication technologies, particularly the Internet, have affected how we, both as individuals and as a society, create, store, and recall information. Internet also provides us with a great opportunity to study memory using transactional large scale data, in a quantitative framework similar to the practice in statistical physics. In this project, we make use of online data by analysing viewership statistics of Wikipedia articles on aircraft crashes. We study the relation between recent events and past events and particularly focus on understanding memory triggering patterns. We devise a quantitative model that explains the flow of viewership from a current event to past events based on similarity in time, geography, topic, and the hyperlink structure of Wikipedia articles. We show that on average the secondary flow of attention to past events generated by such remembering processes is larger than the primary attention flow to the current event. We are the first to report these cascading effects.

 

Memory Remains: Understanding Collective Memory in the Digital Age
Ruth García-Gavilanes, Anders Mollgaard, Milena Tsvetkova, Taha Yasseri

Source: arxiv.org

We show how the success of deep learning depends not only on mathematics but also on physics: although well-known mathematical theorems guarantee that neural networks can approximate arbitrary functions well, the class of functions of practical interest can be approximated through “cheap learning” with exponentially fewer parameters than generic ones, because they have simplifying properties tracing back to the laws of physics. The exceptional simplicity of physics-based functions hinges on properties such as symmetry, locality, compositionality and polynomial log-probability, and we explore how these properties translate into exceptionally simple neural networks approximating both natural phenomena such as images and abstract representations thereof such as drawings. We further argue that when the statistical process generating the data is of a certain hierarchical form prevalent in physics and machine-learning, a deep neural network can be more efficient than a shallow one. We formalize these claims using information theory and discuss the relation to renormalization group procedures. Various “no-flattening theorems” show when these efficient deep networks cannot be accurately approximated by shallow ones without efficiency loss – even for linear networks.

 

Why does deep and cheap learning work so well?
Henry W. Lin, Max Tegmark

Source: arxiv.org

We study the dynamic network of real world person-to-person interactions between approximately 1,000 individuals with 5-min resolution across several months. There is currently no coherent theoretical framework for summarizing the tens of thousands of interactions per day in this complex network, but here we show that at the right temporal resolution, social groups can be identified directly. We outline and validate a framework that enables us to study the statistical properties of individual social events as well as series of meetings across weeks and months. Representing the dynamic network as sequences of such meetings reduces the complexity of the system dramatically. We illustrate the usefulness of the framework by investigating the predictability of human social activity.

 

Fundamental structures of dynamic social networks
Vedran Sekara, Arkadiusz Stopczynski, and Sune Lehmann

PNAS vol. 113 no. 36

Source: www.pnas.org

Spontaneous synchronization has long served as a paradigm for behavioral uniformity that can emerge from interactions in complex systems. When the interacting entities are identical and their coupling patterns are also identical, the complete synchronization of the entire network is the state inheriting the system symmetry. As in other systems subject to symmetry breaking, such symmetric states are not always stable. Here, we report on the discovery of the converse of symmetry breaking—the scenario in which complete synchronization is not stable for identically coupled identical oscillators but becomes stable when, and only when, the oscillator parameters are judiciously tuned to nonidentical values, thereby breaking the system symmetry to preserve the state symmetry. Aside from demonstrating that diversity can facilitate and even be required for uniformity and consensus, this suggests a mechanism for convergent forms of pattern formation in which initially asymmetric patterns evolve into symmetric ones.

 

Symmetric States Requiring System Asymmetry
Takashi Nishikawa and Adilson E. Motter
Phys. Rev. Lett. 117, 114101

Source: journals.aps.org

See Also: Synopsis: Diversity Breeds Conformity