September 01 – Detailed Program

CD-MAKE Keynote - Whiteboxing machine learning

The deployment of AI systems based on machine learning (ML) in real world scenarios faces a number of challenges due to its black box nature. The GDPR right to an explanation has given rise to frantic attempts to develop and design self-explanatory systems, meant to help people understand their decisions and behaviour. In this keynote I will explain (pun intended) why meaningful explanations require keen attention to the proxies used in ML research design. Once those confronted with the decisions or behaviour of ML systems have a better understanding of the pragmatic choices that must be made to allow a machine to learn, it will become easier to foresee what ML systems can and cannot do. Whiteboxing ML should focus on the proxies that stand for real world events, actions and states of affairs, highlighting that a proxy (dataset, variable, model) is not what it stands for.

more...

Standing Still is Not An Option: Alternative Baselines for Attainable Utility Preservation

Enhancing Trust in Machine Learning Systems by Formal Methods

Sustainability Effects of Robust and Resilient Artificial Intelligence

The Split Matters: Flat Minima Methods for Improving the Performance of GNNs

Using Machine Learning to Generate an ESG Dictionary

A Quic(k) Security Overview: A Literature Research on Implemented Security Recommendations

Secure Multi-User Contract Certificate Management for ISO 15118-20 Using Hardware Identities

Simulated environment for multiparty quantum digital signature across the network

An Analysis of Stack Exchange Questions: Identifying Challenges in Software Design and Development with a Focus on Data Privacy and Data Protection

JITScanner: Just-in-Time Executable Page Check in the Linux Operating System

Exploiting Digital Twin technology for Cybersecurity Monitoring in Smart Grids

Program Characterization for Software Exploitation Detection

Obfuscated Mobile Malware Detection by means of Dynamic Analysis and Explainable Deep Learning

Probabilistic framework based on Deep Learning for differentiating ultrasound movie view planes

Let me think! Investigating the effect of explanations feeding doubts about the AI advice

Memorization of Named Entities in Fine-tuned BERT Models

Event and Entity Extraction from Generated Video Captions

Fine-Tuning Language Models for Scientific Writing Support

Experiences with Secure Pipelines in Highly Regulated Environments

Creating a Decryption Proof Verifier for the Estonian Internet Voting System

Attack on "A Privacy-Preserving Online Ride-Hailing System Without Involving a Third Trusted Server"

Exploring NFT Validation through Digital Watermarking