News

FIT team helps build a unique dialect map

A team from the Faculty of Information Technology at the BUT, led by Martin Karafiát, is involved in a unique dialect mapping project. In cooperation with the Academy of Sciences of the Czech Republic and Palacký University in Olomouc, they are creating a website where you can select a region of the Czech Republic and listen to the dialects characteristic of a given place. In addition, the project team is categorising the recordings, which go back to the 1950s, according to various criteria, such as the themes of the narratives.


Researchers from the Dialectology Department of the Institute of Czech Language of the Czech Academy of Sciences have long been trying to map and preserve the various dialects across the Czech Republic. In 2023, they have enlisted the help of experts from the Speech@FIT group, who are currently working on creating a system that would be able to identify the dialect. And also create an automatic transcription of the recordings. "Our speech group has had great success in the areas of language identification, speaker identification and speech transcription. So the primary idea is to bring these areas together, work with unique data and create a system that will be able to automatically transcribe the audio data, which will be a huge help to researchers at the Academy of Sciences. Especially because the data is specific and the classic transcribers from Google or Microsoft fail," explains Martin Karafiát from FIT BUT.


You can read more about this project in our article here.

[img]

How to defend against deepfakes? Kamil Malinka from FIT presented the risks of AI abuse

On Wednesday 31 January, the Economic Committee of the Chamber of Deputies held a seminar on the security threats of artificial intelligence. Kamil Malinka from FIT BUT presented his expert view to a wide audience from the state, commercial and academic spheres as well as representatives of defrauded citizens at the invitation of MP and chairman of the subcommittee on consumer protection Patrik Nacher. 

He spoke about deep fake scams. Artificial intelligence is constantly increasing their sophistication. "These are attacks on biometric systems, identity theft, hate propaganda and more. Artificial intelligence gives attackers "powerful new tools" against which we are relatively powerless", Kamil Malinka explains. Voice deepfakes are already at such a high level that it is impossible to distinguish an artificial voice from a real one by ear. With a human voice, an attacker calls and demands something. It may seem like someone you know is calling from a known number, but in reality it is a complete stranger from a different number and you won't have the slightest suspicion that it is a fake. What's more, this is how a robot programmed for this purpose can communicate with you.  The success rate of such calls increases with how much or how little one concentrates on them. When attention is fragmented and under pressure, people are much more likely to fall for a scam.



Kamil Malinka is working on the issue of deep fakes and their impact on security as part of the Security@FIT research group. Together with another group of researchers from Speech@FIT and Phonexia, they are collaborating on cybersecurity research as part of a Home Office challenge. The aim is to develop tools that can reliably identify man-made recordings.

[img]

Paint like Pollock. Mobile app by FIT BUT students awakens the inner artist in children and adults

Brushes, paints, palette, canvas - all these are usually considered basic necessities for artistic expression. Designed for all art lovers, regardless of age, the Pollock Artify app offers the opportunity to unleash your creativity with a tool that is always at hand - your mobile phone. Its movements are transformed into brush movements on the canvas in the style of American abstract painter Jackson Pollock.  It was created by four students of the Faculty of Information Technology, Peter Zdravecký, Slavomír Svorada, Jakub Zaukolec and Jozef Čásar. 

The idea to connect the world of aesthetics and technology came from a reflection on human values. "In addition to the basic needs in life, humans need to perceive aesthetic values, which include art. Pollock Artify allows you to create your own art and share it with others," explains Peter Zdravecký, one of the members of the implementation team. 


More about the Pollock Artify app here.

[img]

Professor Laura Pozzi from USI Logano at FIT: XPAT: A Parametrizable Template for Approximate Logic Synthesis

From 14 to 15 February we will welcome Professor Laura Pozzi from the Swiss University USI Lugano to our faculty. She will start her program at FIT with a dissertation review of Ing. Jiří Matyáš: The use of formal methods in approximate calculus. 

From 2 pm onwards, she will then give a talk in room E104 on XPAT: A Parametrizable Template for Approximate Logic Synthesis, in which she will discuss her research work on approximate computation, a paradigm for inexact computation with the aim of saving energy, area and/or delay compared to exact computation when the resulting error is bounded. The full abstract of the talk is below.

There will be an opportunity for any discussion after the talk and also on the morning of February 15.

Abstract: In this talk, I will discuss my research work in the field of Approximate Computing, which is a paradigm for computing inexactly in order to save energy, area, and/or delay when compared to exact computation, as long as the incurred error is limited. After providing a brief introduction to the field, I will focus in particular on a method we have recently devised, called XPAT, to derive approXimate circuits through a PArametrical Template by the use of satisfiability techniques. The method takes in input an exact circuit and an error threshold that a user is prepared to tolerate, and in output it generates an approximate circuit which is guaranteed to never differ from the exact by more than the given error threshold. The method's key idea is to fit the desired approximate circuit onto a parametrical template, and to let an SMT solver shape the final solution by choosing the values of the parameters which will make the error constraint satisfiable.

[img]

Voice deepfakes can't be detected by humans or security systems. Attacks on the rise

Spreading alarm messages or disclosing confidential company or bank data. Artificial intelligence is developing rapidly and almost anyone can create deepfake voice recordings at home and in high quality. Neither humans nor biometric systems can reliably distinguish artificial speech from real speech. Researchers from FIT BUT together with commercial system developers now want to design more reliable testing and more accurate detection of deepfakes. They are responding to a call from the Ministry of the Interior.

Anton Firc from FIT BUT first started to address the issue of deepfakes in his master's thesis, in which he investigated the resistance of voice biometrics to deepfakes. The same issue was then followed up by Daniel Prudky's research, which sent voice messages to 31 respondents and investigated their ability to detect deepfakes in ordinary conversation. "People were told a cover story about the user-friendliness of the voicemails being tested. He included one deepfake recording in the test conversations and monitored the respondents' reactions. The results showed that none of them experienced a fraudulent deepfake message," Firc explains.

However, in the same experiment, when respondents were told that one of the voicemails was a fake, they were able to identify it with almost 80% accuracy. "But the research showed that although a deepfake recording is easily identifiable among real ones, no one can detect it in a normal conversation," Firc adds. Part of the reason for this, he says, is that the interviewer didn't expect it in the context, and that's what the creators of deepfake recordings can exploit in reality.


Complete article here.

[img]

Page:

Back to top