The Next Best-Seller Could Be Written by AI: What We Learned at Madrid.AI

City.AI is a global movement organizing events and meetings around the world. Their aim is to share experiences, knowledge, trends, and participate in workshops on Artificial Intelligence (AI). The movement held an event on June 18th in Madrid, organized by Madrid.AI, the Spanish branch of City.AI.

In this meeting, different experts across a wide range of Artificial Intelligence fields shared their most recent work, studies, and analysis on how AI is progressing in some specific areas.

Consequently, this 5th Madrid.AI event covered topics related to the cognitive applications of AI, some Machine Learning case studies, and content production thanks to Deep Learning.

María Borbonés, from IBM Watson Cloud, explained some projects they are developing at her company, the technology they use, and their fields of research.

Using image recognition through a dataset from videos (with applications in security, the fashion industry, among others), data processing (for candidates recruiting at human resources departments), and Natural Language Processing (applied in social networks, personal assistants, psychology…) were some of the fields that María explained as the most promising, with a wide range of applications in many sectors. But, as María stated, “we are not looking to substitute human capacities, but rather to enhance them.”

Attendants at Madrid.AI had the chance to listen to more case studies thanks to Javier Placer, who explained the different phases of an end-to-end Machine Learning project, and how it is being applied across different fields.

Some of the cases that Javier shared are how an African startup predicts the churn rate among its customers, how a Dutch company predicts when a confidentiality contract is needed, and the main requirements of Deep Learning, its applications, and limitations.

Álvaro Barbero focused his talk on Neurowriter, an open-source tool working tech created with Keras and Tensorflow able to produce literary content following a given style. For instance, after introducing the main characteristics of J.K Rowling’s style, Álvaro “wrote” an additional chapter of Harry Potter using the tool, which he shared with the attendants. He has also created a text following Don Quixote’s style, developed cocktail recipes, and created movie titles. This progress is only possible due to residual networks, BPE, bayesian optimization, or bidirectional LSTM.

The talks were followed by Q&As where the attendants could ask experts questions and share their thoughts.

The Madrid.AI community meets every 3 months, and in this 5th edition (sponsored by Terminus7, just like the previous editions) everybody met for a slice of pizza afterward and proposed different styles and genres for the next AI-generated novel.

If you are curious to see all that Artificial Intelligence can do for your business, contact us.