<italic>Computer</italic> Highlights Society Magazines
Summary form only: Summaries of articles presented in this issue of the publication.
50 & 25 Years Ago
Summary form only: Summaries of articles presented in this issue of the publication.
The Influencer
In this message, we ponder whether social media is the dominant influencer of our times.
Also in this Issue
Summary form only: Summaries of articles presented in this issue of the publication.
Assured Autonomy, Artificial Intelligence, and Machine Learning: A Roundtable Discussion
This report summarizes a roundtable panel discussion held at the Second Annual IEEE Workshop on Assured Autonomy, AI, and Machine Learning. Eight expert panelists discussed ways to ensure that artificial intelligence and machine learning systems are safe.
Trustworthy AI—Part III
- Riccardo Mariani,
- Francesca Rossi,
- Rita Cucchiara,
- Marco Pavone,
- Barnaby Simkin,
- Ansgar Koene,
- Jochen Papenbrock
The final part of our Trustworthy AI special issue features four articles on AI security, reliability, trust, and AI trustworthiness as a whole.
Machine Learning Security Against Data Poisoning: Are We There Yet?
Poisoning attacks compromise the training data utilized to train machine learning (ML) models, diminishing their overall performance, manipulating predictions on specific test samples, and implanting backdoors. This article thoughtfully explores these ...
Trustworthy AI-Based Personalized Insulin Recommender for Elderly People Who Have Type-2 Diabetes
We propose TRAINER, a TRustworthy Artificial Intelligence-based iNsulin recommendER for elderly individuals with type 2 diabetes, ensuring reliability and trust in insulin dosage recommendations. TRAINER exemplifies this trustworthiness and addresses such ...
Human-Centered Trustworthy Framework: A Human–Computer Interaction Perspective
The proposed framework (Human-Centered Trustworthy Framework) provides a novel human–computer interaction approach to incorporate positive and meaningful trustful user experiences in the system design process. It helps to illustrate potential users' trust ...
An AI Harms and Governance Framework for Trustworthy AI
Many guidelines have been written and some frameworks proposed for the development of trustworthy artificial intelligence (AI), but a common concern is the lack of precision in definitions, which can make application difficult. This article proposes a ...
Distributed AI in Zero-Touch Provisioning for Edge Networks: Challenges and Research Directions
- Abhishek Hazra,
- Andrea Morichetta,
- Ilir Murturi,
- Lauri Lovén,
- Chinmaya Kumar Dehury,
- Víctor Casamayor Pujol,
- Praveen Kumar Donta,
- Schahram Dustdar
This article combines Distributed Artificial Intelligence (DAI) with zero-touch provisioning (ZTP) for edge networks. Several advantages are also highlighted that come with incorporating DAI into ZTP in the context of edge networks. Further, we draw ...
Learn to Unlearn: Insights Into Machine Unlearning
This article presents a comprehensive review of recent machine unlearning techniques, verification mechanisms, and potential attacks. We highlight emerging challenges and prospective research directions, aiming to provide valuable resources for ...
Video Surveillance and Privacy: A Solvable Paradox?
Through experiments on action recognition and natural language description, we show that the paradox of surveillance and privacy can be solved by artificial intelligence and that respect for human rights is not a chimera.
Making Large Language Models More Reliable and Beneficial: Taking ChatGPT as a Case Study
This article suggests practical ways to make large language models more reliable and beneficial by taking ChatGPT as a case study. Specifically, we describe ChatGPT’s workflow and promised services and highlight the perils requiring the immediate ...
Security Advantages and Challenges of 3D Heterogeneous Integration
- Yuntao Liu,
- Daniel Xing,
- Isaac McDaniel,
- Olsan Ozbay,
- Abir Akib,
- Mumtahina Islam Sukanya,
- Sanjay Rekhi,
- Ankur Srivastava,
- James Bret Michael
Three-dimensional heterogeneous integration offers compelling opportunities to enhance the security and trust in the current semiconductor chain while new attack surfaces may emerge.
How Emerging Memories Extend Battery Life
Energy consumption is an issue with many connected digital products. Resolving energy efficiency issues and putting more memory in less die space create opportunities to use new nonvolatile memories for code storage and cache memory.
Eight Sustainable Practices for Digital Activity Development: Drivers and Barriers in International Higher Education Collaboration
Short-term technology-enhanced activities of international collaboration conducted in virtual, hybrid, or blended modalities can attract a greater number of students. We selected eight sustainable practices that higher education institutions could adopt ...
Safety and Artificial Intelligence in Cyberphysical Systems
Safe artificial intelligence/machine learning (AI/ML)-enabled systems require design methodologies that accommodate for failures in AI/ML decisions. Certification of AI/ML systems and components will increase trust and accelerate adoption, deployment, and ...
Large Language Models and Generative AI, Oh My!
Artificial intelligence (AI) has a miserable reputation regarding promises made and results delivered. Last century, there were many AI winters. The hype surrounding large language models and generative AI could join this trend, we do not hope.