II. Pedia Sapiens: A Planetary Progeny Comes to Her/His Own Actual Factual Knowledge

1. Earthificial Cumulative Cognizance: AI Large Language Models Learn Much Like a Child

Li, Qing, et al. Progress and Opportunities of Foundation Models in Bioinformatics. arXiv:2402.04286. Chinese University of Hong Kong and BioMap, Beijing computer scientists provide a wide-ranging perspective on this mid 2020s synthesis of a Bioinformatic approach, whose journal goes back to 1985, and these novel AI neural net, large language models as they become amenable.

Bioinformatics has witnessed a paradigm shift with the increasing integration of artificial intelligence (AI) and the adoption of foundation models (FMs). These AI techniques have addressed prior issues in bioinformatics such as scarce annotations and of data noise. FMs are adept at handling large-scale, unlabeled data, which has allowed them to achieve notable results in downstream validation tasks. The primary goal of this survey is to conduct a systematic investigation and summary of FMs in bioinformatics, tracing their evolution, current research status, and the methodologies employed. Finally, we outline potential development paths and strategies for FMs in future biological research. (Excerpt)

Lin, Henry and Max Tegmark. Why does Deep and Cheap Learning Work so Well?. arXiv:1608.08225. The Harvard and MIT polymaths review the recent successes of these neural net, multiscale, algorithmic operations (definitions vary) from a statistical physics context such as renormalization groups and symmetric topologies. (Intelligent Evolution)

Liu, Weibo, et al. A Survey of Deep Neural Network Architectures and their Applications. Neurocomputing. 234/11, 2017. As the Abstract cites, Brunel University, London, Xiamen University, Yangzhou University, and King Abdulaziz University, Jeddah, computer engineers provide a wide-ranging tutorial on these increasingly useful cognitive methods.

Since the proposal of a fast learning algorithm for deep belief networks in 2006, the deep learning techniques have drawn ever-increasing research interests because of their inherent capability of overcoming the drawback of traditional algorithms dependent on hand-designed features. Deep learning approaches have also been found to be suitable for big data analysis with successful applications to computer vision, pattern recognition, speech recognition, natural language processing, and recommendation systems. In this paper, we discuss some widely-used deep learning architectures and their practical applications. An up-to-date overview is provided on four deep learning architectures, namely, autoencoder, convolutional neural network, deep belief network, and restricted Boltzmann machine. Different types of deep neural networks are surveyed and recent progresses are summarized. (Abstract)

Liu, Ziming, et al. KAN: Kolmogorov-Arnold Networks.. arXiv:2404.19756. MIT, Caltech and Northeastern University cognitive scholars including Max Tegmark draw on these companion mathematical theories to gin up a new, improved complementary version for the already capable artificial neural nets. See also Novel Architecture Makes Neural Networks More Understandable by Steve Nadis in Quanta for (September 11, 2024) for a good review article.

Inspired by the Kolmogorov-Arnold representation theorem, we propose Kolmogorov-Arnold Networks (KANs) as alternatives to Multi-Layer Perceptrons (MLPs). While MLPs have fixed activation functions on nodes ("neurons"), KANs have learnable functions on edges ("weights"). We show that this seemingly simple change makes KANs outperform MLPs in terms of accuracy and interpretability. Through examples in mathematics and physics, KANs are shown to be useful collaborators helping scientists (re)discover mathematical and physical laws. In summary, KANs open opportunities for improving today's deep learning models. (Excerpt)

Inspired by the Kolmogorov-Arnold representation theorem, we propose Kolmogorov-Arnold Networks (KANs) as promising alternatives to Multi-Layer Perceptrons (MLPs). While MLPs have fixed activation functions on nodes ("neurons"), KANs have learnable activation functions on edges ("weights") by which they can outperform in terms of accuracy and interpretability. Through two examples, KANs are shown to be useful collaborators helping scientists (re)discover mathematical and physical laws. In summary, KANs are promising alternatives for MLPs, opening opportunities for further improving today's deep learning models which rely heavily on MLPs. (S, Nadis)

Lucie-Smith, Luisa, et al. Machine Learning Cosmological Structure Formation. arXiv:1802.04271. We cite this entry by University College London astrophysicists including Hiranya Peiris as an example of the widest range that a new cerebral-based artificial intelligence methods can be applied. If to reflect, whom is this person/sapiensphere prodigy to so proceed as the universe’s way of achieving its own self-quantified description?

Maheswaranathan, Niru, et al. Universality and Individuality in Neural Dynamics across Large Populations of Recurrent Networks. arXiv:1907.08549. By virtue of the latest sophistications, Google Brain and Stanford University AI researchers are able to discern and report “representational similarities” between “biological and artificial networks.” These qualities are then seen in effect across an array of personal and communal affinities.

Manyika, James, ed. AI & Society. Daedulus. Spring, 2022. A timely, dedicated survey with entries like If We Succeed by Stuart Russell, A Golden Decade of Deep Learning by Jeffrey Dean, Language & Coding Creativity by Ermira Murati, and Signs Taken for Wonders: AI. Art & the Matter of Race by Michele Elam.

AI is transforming our relationships with technology and with others, our senses of self, as well as our approaches to health care, banking, democracy, and the courts. But while AI in its many forms has become ubiquitous and its benefits to society and the individual have grown, its impacts are varied. Concerns about its unintended effects and misuses have become paramount in conversations about the successful integration of AI in society. This volume explores the many facets of artificial intelligence: its technology, its potential futures, its effects on labor and the economy, its relationship with inequalities, its role in law and governance, its challenges to national security, and what it says about us as humans. (Issue review)

Manzalino, Antonio. Complex Deep Learning with Quantum Optics. Quantum Reports. 1/1, 2019. In this new MDPI online journal, a senior manager in the Innovation Dept. of Telecom Italia Mobile (TIM), bio below, advances the frontiers of this current assimilation of a lively quantum cosmos with human neural net cognizance. See also, e.g., a cited reference, Quantum Fields as Deep Learning, by Jae-Weon Lee at arXiv:1708.07408. While a prior physics mindset worries over an opaque strangeness, into these later 2010s, via instant global collaborations, a profound new understanding and treatment becomes possible.

The rapid push towards telecommunications infrastructures such as 5G capacity and the Internet drives a strong interest for artificial intelligence (AI) methods, systems, and networks. Processing big data to infer patterns at high speeds with low power consumption is a central technological challenge. Today, an emerging research field rooted in quantum optics along with deep neural networks (DNNs) and nanophotonics are cross-informing each other. This paper elaborates on these topics and proposes a theoretical architecture for a Complex DNN made from programmable metasurfaces. An example is provided which shows a correspondence between the equivariance of convolutional neural networks and the invariance principle of gauge transformations. (Abstract)

Antonio Manzalini received the M. Sc. Degree in Electronic Engineering from the Politecnico of Turin and the Ph.D on Computer Science and Networks from Télécom SudParis and Université Pierre & Marie Curie – Sorbonne. His results have been published in more than 130 of technical papers. His interests in TIM concern SDN and NFV, Cloud vs Multi-Access Edge Computing for 5G, Future Internet and Quantum Communications.

Marchetti, Tomasso, et al. An Artificial Neural Network to Discovery Hypervelocity Stars. arXiv:1704.07990. An eight member European astrophysicist team finds this cerebral procedure to be a fruitful way to distill results of the myriad data findings of the Gaia space telescope mission. Once again, we note how such a collaboration may appear as a worldwide sapiensphere proceeding to learn on her/his own.

The paucity of hypervelocity stars (HVSs) known to date has severely hampered their potential to investigate the stellar population of the Galactic Centre and the Galactic Potential. The first Gaia data release gives an opportunity to increase the current sample. The challenge is of course the disparity between the expected number of hypervelocity stars and that of bound background stars (around 1 in 106). We have applied a novel data mining algorithm based on machine learning techniques, an artificial neural network, to the Tycho-Gaia astrometric solution (TGAS) catalogue. (Abstract excerpt)

Marcus, Gary. The Next Decade in AI: Four Steps Towards Robust Artificial Intelligence. arXiv:2002.06177. The NYU polypsychologist and founder of Robust AI has rightly situated himself as a reality checker and quality control moderator as this hyper-active endeavor moves into every aspect that it can. See also his Rebooting AI: Building Artificial Intelligence We Can Trust 2019 book.

Recent research in artificial intelligence and machine learning has largely emphasized general-purpose learning and ever-larger training sets and more and more compute. In contrast, I propose a hybrid, knowledge-driven, reasoning-based approach, centered around cognitive models, that could provide the substrate for a richer, more robust AI than is currently possible.

Mathuriya, Amrita, et al. CosmoFlow: Using Deep Learning to Learn the Universe at Scale. arXiv:1808.04728. As the brain-based AI revolution proceeds, seventeen authors from Intel, LBNL, Cray and UC Berkeley scope out their neural network applications, as being done everywhere else, across the celestial raiment. Indeed, as this realm becomes similarly amenable, one might get the impression that the whole cosmos is somehow cerebral, or genomic in nature.

Deep learning is a promising tool to determine the physical model that describes our universe. To handle the considerable computational cost of this problem, we present CosmoFlow: a highly scalable deep learning application built on top of the TensorFlow framework. CosmoFlow uses efficient implementations of 3D convolution and pooling primitives, together with improvements in threading for many element-wise operations, to improve training performance on Intel(C) Xeon Phi(TM) processors. We also utilize the Cray PE Machine Learning Plugin for efficient scaling to multiple nodes. To our knowledge, this is the first large-scale science application of the TensorFlow framework at supercomputer scale with fully-synchronous training. (Abstract)

Deep Learning for Cosmology: The nature of dark energy is one of the most exciting and fundamental questions facing scientists today. Dark energy is the unknown force that is driving the accelerated expansion of the universe, and is the subject of several current and future experiments that will survey the sky in multiple wavelengths. We cannot measure dark energy directly - we can only observe the effect it has on the observable universe. The interplay of gravity (pulling matter together) and dark energy (expanding space itself) is encoded in the distribution of matter in the universe today. Cosmologists typically characterize this distribution using statistical measures of the structure of matter – its “clumpiness” - in the form of two- or three-point correlation functions or other reduced statistics. Methods that capture all features in the distribution of matter (such as deep learning networks) could give greater insight into the nature of dark energy. (1)

Mehta, Pankaj and David Schwab. An Exact Mapping between the Variational Renormalization Group and Deep Learning. arXiv:1410.3831. We cite this entry because Boston University and Northwestern University physicists show a common affinity between this intelligent neural net method and dynamic physical qualities. So we muse, could it be imagined that cosmic nature may be in some actual way cerebral in kind, engaged in a grand educative experience. One is reminded of the British physicist James Jeans’ 1930 quote The universe begins to look more like a great thought than like a great machine. See also Machine Learning meets Quantum State Preparation at arXiv:1705.00565.

Deep learning is a broad set of techniques that uses multiple layers of representation to automatically learn relevant features directly from structured data. Recently, such techniques have yielded record-breaking results on a diverse set of difficult machine learning tasks in computer vision, speech recognition, and natural language processing. Despite the enormous success of deep learning, relatively little is understood theoretically about why these techniques are so successful at feature learning and compression. Here, we show that deep learning is intimately related to one of the most important and successful techniques in theoretical physics, the renormalization group (RG). RG is an iterative coarse-graining scheme that allows for the extraction of relevant features (i.e. operators) as a physical system is examined at different length scales. Our results suggests that deep learning algorithms may be employing a generalized RG-like scheme to learn relevant features from data. (Abstract)

Previous 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 Next