What are neural networks<i>not</i>good at? On artificial creativity

Олейник, Антон

doi:10.1177/2053951719839433

Cited by 13 publications

(7 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Deep neural networks with far more parameters than training data can also be reliably trained [46], but it is a long way from the current artificial neural network to produce human-like intelligence [47].…”

Section: Discussionmentioning

confidence: 99%

Image Quality Predictor with Highly Efficient Fully Convolutional Neural Network

Cao

Xin-Lin

Liu

2022

Advances in Multimedia

View full text Add to dashboard Cite

A highly efficient deep fully convolutional neural network (DFCN) for image quality assessment (IQA) is designed in this paper. The DFCN consists of two branches, one scoring local patches and the other estimating the weights of local patches to enhance quality prediction. Then, the DFCN outputs quality score of the whole image with aggregate weighted average pooling. There are no fully connected layers in the DFCN, resulting in far fewer parameters. In addition, the network model utilizes multiscale images as inputs to enrich the extracted distortion information. Furthermore, the parameters of the model are optimized in two steps to reduce the requirement for computing power and the risk of overfitting. The parameters of the shared layers and the quality module are optimized firstly, and then, the parameters of the weight module are optimized with the designed loss function. The extensive experimental results show that the proposed DFCN outperforms other competing IQA methods and has strong generalization ability.

show abstract

Section: Discussionmentioning

confidence: 99%

Image Quality Predictor with Highly Efficient Fully Convolutional Neural Network

Cao

Xin-Lin

Liu

2022

Advances in Multimedia

View full text Add to dashboard Cite

show abstract

“…Devising autonomous systems is almost impossible for many real-world scenarios where humans need to "stay in the loop" to maintain the sociotechnical system's versatility and adaptability in relation to new tasks and environments. Some of the greatest challenges in developing AI systems have to do with bringing contextual meaning and reasoning in relation to real-world situations (Oleinik, 2019). This requires continuous human-AI interactions, a vision beyond "automating the last mile" and superior performance in narrowly defined tasks.…”

Section: Conclusion and Future Research Agendamentioning

confidence: 99%

Artificial intelligence, human intelligence and hybrid intelligence based on mutual augmentation

2022

View full text Add to dashboard Cite

There is little consensus on what artificial intelligence (AI) systems may or may not embrace. Although this may point to multiplicity of interpretations and backgrounds, a lack of conceptual clarity could thwart the development of common ground around the concept among researchers, practitioners and users of AI and pave the way for misinterpretation and abuse of the concept. This article argues that one of the effective ways to delineate the concept of AI is to compare and contrast it with human intelligence. In doing so, the article broaches the unique capabilities of humans and AI in relation to one another (human and machine tacit knowledge), as well as two types of AI systems: one that goes beyond human intelligence and one that is necessarily and inherently tied to it. It finally highlights how humans and AI can augment their capabilities and intelligence through synergistic human–AI interactions (i.e., human-augmented AI and augmented human intelligence), resulting in hybrid intelligence, and concludes with a future-looking research agenda.

show abstract

“…[19] The creativity methods are therefore not suitable for this task, as the aim of this paper is to lay out a foundation for an automated determination of risks. Current information systems are not suitable for a creative performance, but are very capable of finding patterns in huge sets of data [20].…”

Section: Risk Modelling and Knowledge Preservationmentioning

confidence: 99%

Towards a Unified Reliability-Centered Information Logistics Model for Production Assets

Defèr¹,

Schuh²,

Stich³

2020

IFIP Advances in Information and Communication Technology

View full text Add to dashboard Cite

Reliability-centered maintenance for production assets is a well-established concept for the most effective and efficient disposition of maintenance resources. Unfortunately, the approach takes a lot of effort and relies heavily on the knowledge of individuals. Reliability data in Computerized Maintenance Management System (CMMS) is scarce and almost never used well. An automated risk assessment system would have the potential to contribute to the dissemination and effective use of risk information and analysis. The individuality of production setting, however, prevents current systems from being practically relevant for most industries. The presented approach combines ontologies to store and link knowledge, an information logistics model displaying the various information streams, and the Internet of production to take the different user systems and infrastructure layers into account. The provided model of a reference digital shadow for risk information and a detailed information logistics model will help software companies to improve reliability software, standardize and enable assets owners to establish a customized digital shadow for their production networks.

show abstract

What are neural networksnotgood at? On artificial creativity

Cited by 13 publications

References 45 publications

Image Quality Predictor with Highly Efficient Fully Convolutional Neural Network

Image Quality Predictor with Highly Efficient Fully Convolutional Neural Network

Artificial intelligence, human intelligence and hybrid intelligence based on mutual augmentation

Towards a Unified Reliability-Centered Information Logistics Model for Production Assets

Contact Info

Product

Resources

About