1,721,136 research outputs found
Kinship recognition: How far are we from viable solutions in smart environments?
No full textAutomated kinship recognition from face is a relatively recent problem that is mainly studied by the application of Deep Learning techniques. Despite the impact that an accurate kinship recognition algorithm can reach in a controlled environment, its applicability in smart environments is limited due to the degradation of performances. In this study we investigate the limitations of recent approaches that lead to a difficult applicability in a real case use. We present several tests on Siamese Neural Networks (SNN) based on a VGGFace architecture to solve both the kinship-vs-not-kinship recognition and the kind-of-kinship recognition. To perform our tests we used two popular kinship recognition Datasets that are Faces in the Wild and KinFace-II, respectively. To examine the behavior of the SNNs in a real scenario, we applied them, properly trained on the above mentioned datasets, to a popular TV show in which the aim is to discover kinship in a set of people. The weaknesses demonstrated in those tests have confirmed that the recent literature and algorithm to solve the kinship recognition problem are still far to achieve the high performances required in a smart environment
A general model for fair and explainable recommendation in the loan domain
No full textRecommender systems have been widely used in the Financial Services domain and can play a crucial role in personal loan comparison platforms. However, the use of AI in this domain has brought to light many opportunities as well as new ethical and legal risks. The customers can trust the suggestions of these systems only if the recommendation process is Interpretable, Understandable, and Fair for the end-user. Since products offered within the banking sector are usually of an intangible nature, customer trust perception is crucial to maintain a long-standing relationship and ensure customer loyalty. To this end, in this paper, we propose a model for generating natural language and counterfactual explanations for a loan recommender system with the aim of providing fairer and more transparent suggestions
A Pre-Rendering Based Approach to High Detail Reproduction and Navigation of an Archaeological Site
No full textKnowledge Transfer and Crowdsourcing in Cyber-Physical-Social Systems
No full textThe rapid development of cyber-physical systems results in vast amount of heterogeneous data generated every day. To deal with unstructured data and maintain its security and privacy in smart manufacturing, it is necessary to merge social space with cyber-physical systems to develop cyber-physical-social systems. Crowdsourcing and knowledge transfer can be effective approaches to solve problems of manufacturing and product development, such as inviting hackers to break the security bridge to test the efficacy of the measures and handling enormous data generated in cyber-physical-social system. Crowdsourcing is a novel computing paradigm that leverages human effort to tackle computationally difficult issues, whereas knowledge transfer helps complete the new assignment based on quick access to the existing knowledge. As a result, an enhanced annotation for the work may be done at a low cost via suitable knowledge transfer. This paper introduces cyber-physical-social system and highlights the challenges and issues arising from massive data generated over the internet by various sources, and discusses how pattern recognition techniques can be used to identify anomalies or attacks. It also defines the terms knowledge transfer and crowdsourcing and explains how they can be effectively used to solve a problem in cyber-physical-social systems
From Fully Supervised to Blind Digital Anastylosis on DAFNE Dataset
No full textAnastylosis is an archaeological term consisting in a reconstruction technique whereby an artefact is restored using the original architectural elements. Experts can sometimes imply months or years to carry out this task counting on their expertise. Software procedures can represent a valid support but several challenges arise when dealing with practical scenarios. This paper starts from the achievements on DAFNE challenge, with a traditional template matching approach which won the third place at the competition, to arrive to discuss the critical issues that make the unsupervised version, the blind digital anastylosis, a hard problem to solve. A preliminary solution supported by experimental results is presented
Introduction to the special issue on “Biometrics in Smart Cities: Techniques and Applications (BI_SCI)”
No full textFostering secure cross-layer collaborative communications by means of covert channels in MEC environments
No full textRecently, due to unexpected conditions introduced by the COVID-19 outbreak, collaborative tools are widely adopted in almost all sectors of our daily lifestyle. Almost all tools rely mainly on the World Wide Web technologies that, in turn, are built upon the HTTP protocol. The HTTP protocol is considered as the “bricks” of all kind of communications among people/devices that exchange messages with different purposes and meanings. Unfortunately, it is widely used to track and monitor people when using the Internet. This paper exploits the HTTP protocol and try to reverse this negative aspect by designing and implementing a way to help users (and devices) to not disclose too much information when collaborating each other even in an unfriendly environment. A novel steganographic protocol is proposed by using the HTTP “control” messages. The proposed protocol can be easily adopted by devices communicating in a MEC (Mobile Edge Computing) environment where it is important to guarantee the integrity and the confidentiality of all communications, especially messages that give “instructions” to devices or in device-to-device communications. The proposed protocol allows to avoid using complex and computationally demanding cryptographic protocols that are very difficult to be used in devices with limited resources
Waiting for Tactile: Robotic and Virtual Experiences in the Fog
No full textSocial robots adopt an emotional touch to interact with users inducing and transmitting humanlike emotions. Natural interaction with humans needs to be in real time and well grounded on the full availability of information on the environment. These robots base their way of communicating on direct interaction (touch, listening, view), supported by a range of sensors on the surrounding environment that provide a radially central and partial knowledge on it. Over the past few years, social robots have been demonstrated to implement different features, going from biometric applications to the fusion of machine learning environmental information collected on the edge. This article aims at describing the experiences performed and still ongoing and characterizes a simulation environment developed for the social robot Pepper that aims to foresee the new scenarios and benefits that tactile connectivity will enable
A general aspect-term-extraction model for multi-criteria recommendations
No full textIn recent years, increasingly large quantities of user reviews have been made available by several e-commerce platforms. This content is very useful for recommender systems (RSs), since it reflects the users' opinion of the items regarding several aspects. In fact, they are especially valuable for RSs that are able to exploit multi-faceted user ratings. However, extracting aspect-based ratings from unstructured text is not a trivial task. Deep Learning models for aspect extraction have proven to be effective, but they need to be trained on large quantities of domain-specific data, which are not always available. In this paper, we explore the possibility of transferring knowledge across domains for automatically extracting aspects from user reviews, and its implications in terms of recommendation accuracy. We performed different experiments with several Deep Learning-based Aspect Term Extraction (ATE) techniques and Multi-Criteria recommendation algorithms. Results show that our framework is able to improve recommendation accuracy compared to several baselines based on single-criteria recommendation, despite the fact that no labeled data in the target domain was used when training the ATE model
- …
