1,721,204 research outputs found
Journal of Future Robot Life
No full textWhat will robots be like ten, twenty and more years from now? What will they be able to accomplish? How will human–robot relationships have advanced? What place in society will be occupied by robots? These are just some of the questions which will be debated in the pages of this new publication – the Journal of Future Robot Life
The impact of AI on the musical world: will musicians be obsolete?
Full text linkArtificial intelligence (AI) is going through a period of renewed interest and success thanks to the rise of neural networks, the staple of the so-called deep learning. Creating a computer program capable of writing believable music has been tried since the 1960s, with lackluster results. Composing music seemed something beyond the potential of machines, but recent developments in the field are challenging this conception. This article will explore the latest developments falling at the intersection between artificial intelligence and music and then investigate what possible impact such new technologies may have on the musical world, from a technical as well as an aesthetic standpoint, trying to demystify some common misconceptions and worries
Adaptive Video Transmissions over the Internet: an Experimental Study
No full textSeveral distributed multimedia applications, such as on-line games, video streaming and IF-telephony, require interactive operations between who provides the service (the sender) and who receives the service (the receiver). Considering video transmissions, interactive operations are well performed only if the receiver plays out video frames continuously and with short delay with respect to their transmission time. Unfortunately, this is not trivial especially over networks, as the Internet, whose traffic conditions greatly affect the packets transfer delay. This poses significant problems. For instance, it can happen that a frame f(i), supposed to be played out at time t(i), cannot be played out until time t(i)(double dagger). We denote as Video Time Difference, the difference t(i)(double dagger)-t(i) and in this paper we propose a mechanism that keeps tight and constant this video time difference. This is performed by slightly modifying the QoS of the delivered video. The proposed mechanism has been evaluated through simulations over real Internet traffic conditions and it is shown to be effective in supporting distributed multimedia interactive applications
Informing Clients through Multimedia Communications: An approach to provide interactivity
Full text linkOne of the key problems in informing clients through multimedia streaming applications over the Internet is to customize the stream of information according to the client’s requests. This is achievable only if client and server can interact along the application lifetime, which is possible only if the communication system supports the rigid timing constraints imposed by these interactive applications on their traffic. In the Internet scenario, these applications are very difficult to support, as the Internet provides a best-effort service to the traffic it carries, which means that the Internet does not make any promises about the end-to-end delay for an individual packet and about the variation of packet delay (network jitter) within a packet stream. These problems are confirmed by several experiments we performed over the Internet, which highlight that interactive applications achieve a quality that is frustrating. The contribution of this paper is the proposal of a novel mechanism to support interactive multimedia streaming applications over the Internet. Our mechanism adapts the multimedia stream transmission to the network conditions, by intentionally and slightly acting on the video QoS. Our mechanism has been validated through severalexperiments performed over the Internet. Results confirm that the supported interactive applications achieve a satisfactory quality and the user perceives a video quality only slightly affected by the QoS modification introduced by our mechanism
Interactive MPEG Video Streaming over IP-Networks: A Performance Report
No full textThe popularity of interactive video streaming applications have pushed researchers to propose mechanisms for supporting these applications over the Internet. Several studies showed that interactive operations are well supported if the end-to-end delay, experienced by the application traffic, is kept lower than a pre-defined threshold. Recently, it has been proposed a new approach that acts on the video QoS (by dropping frame) in order to provide interactive features. The mechanism has been tested with Motion JPEG videos. The contribution of this paper is to evaluate the mechanism with MPEG videos, in order to investigate the behavior of the mechanism when the video is encoded with one of the more popular inter-frame techniques. Further, since the mechanism acts on the video QoS, we also present a QoS evaluation of the perceived video play out quality
Design and Analysis of a mechanism for supporting Interactive Video Applications over the Internet
No full textInteractive video streaming applications are becoming more popular also in the Internet environment and studies showed that interactive operations are well supported if the end-to-end delay, experienced by the application traffic, is kept lower than a pre-defined, and application dependent, threshold. In this paper we propose a new approach that acts on the video QoS in order to keep the end-to-end delay within the acceptable threshold. Our mechanism has been evaluated through several simulations and results obtained show that it is well suited for supporting interactive video streaming applications over the Internet
I limiti e le implicazioni di una predizione automatica
No full textNegli ultimi anni il Machine learning ha fatto progressi sorprendenti grazie all’analisi dell’enorme mole di dati a disposizione. Ma, sostengono gli autori, c’è ancora tanta strada da fare prima di vedere una macchina prendere in autonomia decisioni corrette in situazioni delicate.
Le dimensioni problematiche da risolvere sono diverse sia di natura tecnica che etica. La soluzione è certamente combinare human e machine intelligence e usare il deep learning per ottimizzare il l’analisi dei dati
A Preface to New Frontiers for Entertainment Computing - the First IFIP Entertainment Computing Symposium - 20th IFIP World Computer Congress
No full textThis is a preface to the Proceedings of the First IFIP Entertainment Computing Symposium - 20th IFIP World Computer Congress, held in Milan September 200
Fashion, Digital Technologies, and AI. Is the 2020 Pandemic Really Driving a Paradigm Shift?
Full text linkIs the COVID-19 pandemic going to force the fashion industry to rethink herself and push it to embrace digital technologies more massively than before? The answer is most likely “yes”, but the question is somewhat ill-posed. In fact, the fashion world, especially haute couture, has always been very keen to innovation and to digital technology. Even before the current situation, there have been experiments that encompass every part of the fashion ecosystem, including smarter supply chain and manufacturing, design of new materials, new ways of presenting fashion with digitally augmented shows. While other businesses are hardly learning that digital is the way to go, the fashion world seems to have found this insight a long time ago and has been a fertile field for digital applications for a long time. For example, the commercial model has already shifted from being centered around retailers to being heavily reliant on online shopping. Not only this, but we are also seeing an increasing number of so-called digital native fashion brands, that is brands designed from the ground up to be entities of the digital world. This new way of selling fashion has been leveraging big data for some years now. Nonetheless, the abrupt change in our life dictated by the global advent of COVID-19, with the measures taken to mitigate it, like quarantine for example, is most certainly having an further effect on this industry, at all levels, from haute couture to fast fashion, from big brands to small ones. Some few examples include big fashion shows, where dazzling set pieces and parties are no longer possible, replaced by internet live streams. Even big fairs are now hosted as online events, with many brands launching digital applications that allow customers to try clothes virtually. All this considered, while it is certainly true that what happened in 2020 has had the primary effect of relegating retail stores almost to mere warehouses, with the catastrophic possibility they can even disappear in the foreseeable future, yet we believe that the correct question to ask is whether this phenomenon has just started now or has simply accelerated with the onset of the COVID-19 pandemic.In this paper, we favor this second hypothesis, and maintain that the current shift in the fashion industry practices and priorities follow a trend started may years ago, that the spread of the virus has only emphasized
On the probabilistic mind of a robot
Full text linkIn this article, we discuss what differentiates an artificial mind from a human one, during the process of making
a choice. We do this without any intention to debunk famous arguments according to which there are aspects of human
consciousness and expertise that cannot be simulated by artificial humanlike cognitive entities (like a robot, for example).
We will first put in evidence that artificial minds, built on top of Deep Neural Network (DNN) technologies, are probabilistic
in nature, and follow a very clear line of reasoning. Simply told, their reasoning style can be assimilated to a process that
starts from a bunch of example data and learns to point to the most likely output, where the meaning of likely, here, is
neither vague or fuzzy, but it obeys well-known probability theories. Nonetheless, as such, choices could be made by those
intelligent entities that fail to pass human plausibility criteria, even if those chosen are those with high probability values.
We will provide an (obvious) explanation for this apparent paradox, and we will demonstrate that an artificial mind based
on a DNN can be driven to translate probabilities into choices that humans judge as plausible. Finally, we will show how
an artificial probabilistic mind can be made to learn from its errors, till the point where it exhibits a cognitive behavior
comparable to that of a human being
- …
