I’m an information designer with a strong passion for
internet geographies. I’m particularly interested in questioning the ethics of
machine learning systemsand to create my own
datasets. I occasionally develop my own tools to do
graphic design, or to gather information from the
internet. Apart from that, i photograph weird food and i share music with my brother on a weekly basis. If you are interested in this random nonsense, feel free to write me digitally or physically.
This book wants to bring insights to the Caporalato system and an overview to both the main agricultural activities in Italy and to how this topic is being discussed. Caporalato, a seasonal and regional system targeted towards hiring farm labour at very low wages, is active all year around, but very hard to track and to prevent it. Based on migrant interviews, satellite imagery analysis and twitter datas, a structure similar to an atlas makes this book act like a sort of manifesto suggesting possible ways to study the topic.
In summer 2017 the flow of migrants to Italy suddenly dropped by 87%, while at the same time different news sources started speaking about a dodgy deal between Italy and a militia situated in Sabratha, North of Libya. Triggered by a Facebook post depicting a militian with Italian supplies, this website reconstructs, through social media feeds, the clashes and conflicts that took place in Sabratha during the end of September. Furthermore, cartography is used to highlight all fightings between multiple parties contesting the ownership of migrants.
In contemporary society, the atlas is no longer just a container of maps, views and static illustrations. Its definition has expanded with the evolution of information design and, consequently, we should not currently talk of atlases in the traditional sense, but from the means of distributing analytical information. Very often it happens that when a book is published, the arguments in it are already exceeded, therefore we should try to contextualize and select the contents thoroughly. Designing maps and atlases in the twenty-first century means to completely rethink the process. Consequently, the purpose of the current thesis is to analyse the evolution of this sector, considering the positive and negative aspects, potentialities and problems. Since, with the presence of digital tools and satellites, printed maps began to appear inappropriate in the evolution of map making, the focus shifted on their aesthetic. Here, i want to stress how the designer could have a significant role in the development of new atlases, by creating different reading points. I also want to stress the importance, for designers themselves , to be more even able to code, to optimise their ways of researching and designing, which are essential aspects for the creation of future atlases.
Tay, the chatbot created by Microsoft and released via Twitter in 2016, is an example of the misuse of machine learning systems.
The chatbot, which learned by conversing with other Twitter accounts, started to post offensive tweets and within a day was deleted.
Who taught Tay? How many users were participating, and from which part of the World? By finding a text file on Reddit containing a list of tweets of users who engaged with Tay, it was possible to recreate a map based on the geographies of these teachers. All of it has been achieved by retrieving, for each users the personal location that they shared on Twitter.
This interactive map is part of a larger research aimed at mapping the geographies of machine learning datasets.
The central goal of computer vision systems is to make sense of visual material. This is achieved by feeding large-scale collections of semantically enriched images – where each element in each frame has been explicitly described by a human being – to artificial agents. Based on this ‘reading’ of the source material, Artificial Intelligence learns to discern what our world looks like. Most of the time such datasets are its only teachers, and thus the datasets play a crucial role in machine learning. When studying them carefully, it becomes clear that in general little attention is paid to the quality and the origins of the data which carry – together with information – also unfairness and biases.
The current project focuses on a specific dataset: MS COCO. Created by Microsoft Research, COCO shows many examples of common objects in context and offers ground truth guidelines for many computer vision systems. It has been created by web scraping a large collection of images from Flickr. As a consequence, this gives an implicit US-centred connotation to the resource. Furthermore, when deconstructing the actions that lead to its semantic enrichment, it emerges that a disproportionate number of individuals around the world have been involved as AI’s unconscious teachers.
The project is an attempt to reverse engineer this resource in order to uncover and geolocate all the individuals involved in the creation of COCO. In the form of an interactive database, the project reveals the immense workforce involved in the education of AI and the unbalanced geographies embedded within it.
This website investigates the socio-cultural biases present in machine learning algorithms and the representation of the World that is currently taught to Artificial Intelligence.
The disproportionate and naive use of Machine learning Algorithms risks to emphasise stereotypes present in data. Whenever data reflects biases of the broader society, the learning algorithm captures and learns from these stereotypes. This is a matter of concern with Text Embeddings, a popular and worldwide used framework to represent text as vectors. Embeddings trained on Wikipedia, mainly used in softwares such as Google Home and Google Translate, or even in social media like Facebook and Instagram showcase sexism to a disturbing extent. Having vectors it’s then possible to calculate distances, therefore it appears that a word such as “Nurse” will be much more closer to “She” compared to “He”. On this website we visualize by a given input sentence the distinctions that this algorithm (Word2Vec) makes. Through a lexical analysis it’s possible to compare the subject of the written text with what is attached to it. It will therefore be possible not only to distinguish sexism but also racism (i.e. comparing names from different nationalities to the same profession).
On a daily basis, a script collects images containing faces from Google News and displays them in a webpage.
Using artificial intelligence to define new relationships between illustrators drawing styles. This interactive visualization displays the last 10 years of selected works from the Bologna Children's Book Fair organized using machine learning. To do so, a database with different types of drawing style has been created to specifically teach a machine learning algorithm to distinguish between different types of strokes and textures, gouaches and collages.
By using machine learning, this website collects, on a monthly basis, random images of Bologna (Italy) retrieved from Google Street View. It then displays all the pictures it finds containing graffiti. With the goal of redefining a city's cartography by the often political, often random content of these hidden artefacts, maps will be used to highlite different clusters of reoccurent themes.
This ongoing project (2017 - 2019) focuses on the use of a specific application to draw multiples types of patterns. The countless variations that the tool can produce represent what the Academy of Fine Arts of Bologna constitutes: creativity, multidisciplinarity and mutual inspiration between students and teachers. Specifically, this project wanted to communicate my experience and memories as a student. It aimed to evoke the places where I first met my classmates, the entrance, the corridors, the inner courtyard, as well as the discussions outside the classrooms, the collaborations during the workshops, the continuous exchange of ideas, all represented as a wide system of interdisciplinary interactions.
ARKIVIO is a work in progress that reflects on the temporariness of digital archives. The project analyses the fragility of such archives and the invisible accumulation data. The information that was preserved in an old computer, has been edited and later displayed in a series of volumes containing photos, documents, images, videos, albums and especially conversations on Windows Live Messenger. Those memories are filling the pages of six books, reconstructing part of Francesco's adolescence, from 2008 to 2011, exploring that time when he was between the ages of 15 and 18.
Fascinated by Geocities, the underconstruction website and by the work of the Internet Archive, here there is a collection of the best gifs of the old internet.
Redesigning the way we approach torrent websites. Made with my anonimous brother, this weekly newsletter merges the last movies released online with their plot from Imdb and trailer from Youtube.