No full-text available
To read the full-text of this research,
you can request a copy directly from the author.
Predicting ultra-hard binary compounds via cascaded auto- And hetero-associative neural networks
Abstract
When an auto-associative neural network is trained within any conceptual space, the many rules and schema embodied within that knowledge domain are encoded through its many connection weights and biases. For instance, if such a network's input/output exemplars consist of numerous formulas representing known chemical compounds, subsequent network training will produce connection traces that embody the many implicit rules governing the constraints between constituent elements and their allowed proportionalities. That is to say, the net has gained a statistical `insight' into the patterns of bonding, valence, and charge balance that must be observed in theorizing new chemical compounds. If that network is now made chaotic by random perturbation of its processing elements and connection weights, the resulting network activations will represent the formulas of a wide variety of plausible compounds, many of which may be considered novel from the standpoint of network training. We therefore attain an all-neural search engine for generating a stream of plausible chemical possibilities. Adding subsequent `policing' networks to associate these emerging chemical formulas with various physical and chemical properties, we are able to either filter for sought characteristics or alternatively, assemble expanding materials tabulations of potentially new compounds and their estimated properties. Here, we describe the theory, construction, function, and results of just such an autonomous materials discovery machine, tailored specifically to the search for new ultrahard binary compounds.
... We trained an ANN to emulate canine cancer diagnoses of urine samples based upon GC-MS data presented to it. We used both network skeletonization [35][36][37] and auto-associative filtering [38], followed by anomaly detection. Rule extraction techniques were then applied to this net [35][36][37] to reveal the "logic circuitry" developed within it through both training and the selective pruning of its connection weights. ...
... We used both network skeletonization [35][36][37] and auto-associative filtering [38], followed by anomaly detection. Rule extraction techniques were then applied to this net [35][36][37] to reveal the "logic circuitry" developed within it through both training and the selective pruning of its connection weights. Using this semi-quantitative technique, salient inputs to the model became evident, as well as the interplay between them in determining the ANN prediction of the canine's diagnosis. ...
... On the basis of VOCs collected by headspace SPME and analyzed by GC-MS, the raw ion chromatographs were used to train an ANN to emulate canine cancer diagnoses of urine samples. Both network skeletonization [35][36][37] and auto-associative filtering [38] techniques were used to reveal the most important chromatograph peaks contributing to the canine diagnosis. The network skeletonization approach (Fig 5) indicated the most salient spectral features occurred in the interval from 10 minutes to 14 minutes. ...
Prostate cancer is the second leading cause of cancer death in men in the developed world. A more sensitive and specific detection strategy for lethal prostate cancer beyond serum prostate specific antigen (PSA) population screening is urgently needed. Diagnosis by canine olfaction, using dogs trained to detect cancer by smell, has been shown to be both specific and sensitive. While dogs themselves are impractical as scalable diagnostic sensors, machine olfaction for cancer detection is testable. However, studies bridging the divide between clinical diagnostic techniques, artificial intelligence, and molecular analysis remains difficult due to the significant divide between these disciplines. We tested the clinical feasibility of a cross-disciplinary, integrative approach to early prostate cancer biosensing in urine using trained canine olfaction, volatile organic compound (VOC) analysis by gas chromatography-mass spectroscopy (GC-MS) artificial neural network (ANN)-assisted examination, and microbial profiling in a double-blinded pilot study. Two dogs were trained to detect Gleason 9 prostate cancer in urine collected from biopsy-confirmed patients. Biopsy-negative controls were used to assess canine specificity as prostate cancer biodetectors. Urine samples were simultaneously analyzed for their VOC content in headspace via GC-MS and urinary microbiota content via 16S rDNA Illumina sequencing. In addition, the dogs’ diagnoses were used to train an ANN to detect significant peaks in the GC-MS data. The canine olfaction system was 71% sensitive and between 70–76% specific at detecting Gleason 9 prostate cancer. We have also confirmed VOC differences by GC-MS and microbiota differences by 16S rDNA sequencing between cancer positive and biopsy-negative controls. Furthermore, the trained ANN identified regions of interest in the GC-MS data, informed by the canine diagnoses. Methodology and feasibility are established to inform larger-scale studies using canine olfaction, urinary VOCs, and urinary microbiota profiling to develop machine olfaction diagnostic tools. Scalable multi-disciplinary tools may then be compared to PSA screening for earlier, non-invasive, more specific and sensitive detection of clinically aggressive prostate cancers in urine samples.
... Over the last 20 years, Creativity Machines have been applied to diverse¯elds of endeavor including materials discovery [Thaler, 1998], product development [Plotkin, 2009], music synthesis, and improvisational robotic control [Patrick et al., 2007;Thaler, 2013]. This neural architecture and methodology has been recruited as a general model of consciousness wherein internal network perturbations drive a succession of neuronal activation patterns tantamount to stream of consciousness, while a series of neural modules encoding the system's cumulative, idiosyncratic experience sequentially respond to emulate the parade of intangible emotional states and feelings associated with phenomenal consciousness [Thaler, 1996a[Thaler, , 1997b[Thaler, , 2010[Thaler, , 2012. ...
... Accordingly the perturbed network may \hallucinate" environmental inputs with none being applied, or \imagine" variations upon any applied pattern-based theme. With su±ciently high levels of such weight perturbations the learned constraints within the net are modi¯ed or annulled so as to allow the evolution of confabulatory patterns [Thaler, 1993[Thaler, , 1997a[Thaler, , 1997b[Thaler, , 1998] that markedly di®er from output exemplars the net has already trained upon. The novel patterns thusly produced, are much more than interpolations between training exemplars, as seen in traditional approaches using probabilistic neuron excitations [Haykin, 1994], since the weight disruptions warp the attractor landscape therein, producing mutated and hybridized classes that signi¯cantly di®er from those the net has detected within its training data. ...
... Accordingly, this regime corresponds to a highly e±cient form of ideational search, referred to in the computational creativity literature as a multi-stage [Rowe and Partridge, 1993] wherein constraint relationships between problem attributes are allowed to gently relax in generating new concepts, thus circumventing a scenario in which the monitoring algorithm is swamped with outlandish possibilities. 2 In the patent literature, the monitoring perceptrons within this neural assembly are referred to as \alert associative centers" (AAC) since unlike their internally perturbed brethren, they preserve rather than disrupt their inputÀoutput mappings so as to remain vigilant for useful confabulation patterns. Such AACs may utilize multiple perceptrons variably linking to form associative chains that in high level cognitive terms correspond to an unfolding sequence of impressions concerning the memories and confabulations nucleating within the imagitron [Thaler, 1997a[Thaler, , 1997b[Thaler, , 1998[Thaler, , 2010. Similarly, a plurality of perceptrons, each encoding separate knowledge domains, may form a compound imagitron. ...
By allowing one artificial neural network to govern the synaptic noise injected into another based upon its appraisal of patterns nucleating from such disturbances, a contemplative form of artificial intelligence is formed whose creativity and pattern delivery closely parallels that of human cognition. Drawing upon the theory of fractional Brownian motion, we may derive an equation, verifiable through statistical mechanics, which governs both the novelty and rhythm of pattern turnover within such neural systems. Through this equation, we gain valuable insight into the process of idea formation within the brain, whether that organ is making sense of its environment or itself. In doing so, a relationship between creativity and consciousness is revealed, along with a potential path toward building conscious machine intelligence.
... In the parameters related to the independent sample t-test, where the main values do not show a significant difference in the parameters established for the evaluation, marking a value of 0.262 at the level of equal and nonequal variances. Therefore, the null hypothesis is rejected, Ho = There are no significant differences between the control and experimental groups regarding the impact of virtual environments on the learning of binary inorganic compounds (Thaler, 1998;Hoigné et al., 1985;Wang et al., 2006;Cariati et al., 2006;Rivas et al., 2020). ...
The purpose of the research was to determine the effectiveness of virtual environments in learning binary inorganic compounds at the BGU level of the “San Isidro” Educational Unit. For this purpose, the research had a quantitative approach. Data collection was carried out through knowledge tests on the students (control and experimental group). The results obtained showed that at the control group level a mean of 3.86 was calculated, showing a standard deviation of 1.509 and a mean standard error of 0.302. On the other hand, the calculation showed that, for the experimental group, a mean of 4.46, a standard deviation of 1.576 and a mean standard error of 0.384 were calculated, which indicated that there was no significant difference between both groups when evaluating their knowledge Finally, as a conclusion, the importance of incorporating the “Virtual Quimi Project” project into the teaching activity, whose fundamental purpose is to introduce virtual environments oriented to the learning of binary inorganic compounds. In this sense, the main contribution that the research offers is to ensure that the teaching-learning process of this subject is more affective and effective for the teacher.
... Early on, it became quite evident that the same system, namely, a synaptically perturbed neural network, could provide creative possibilities within any conceptual space. Thus, an associative memory that had been cumulatively exposed to valid chemical formulas produced both novel and plausible theoretical chemistries (Thaler 1998). A perturbed perceptron shown numerous design specifications for consumer products generated plausible new merchandise concepts (Plotkin 2009). ...
Although the definition of the term “creativity” widely varies, recent developments in the field of artificial neural networks (ANN) lend a highly comprehensive model to all accounts of this highly prized cognitive process. From this bottom-up, computational perspective, seminal idea formation results from a noise-driven brain storming session between at least two neural assemblies. Ongoing disturbances both to and within such nets serve to drive a sequence of neuron activation patterns that is tantamount to stream of consciousness. At sufficiently intense disturbance levels, memories and their interrelationships degrade into false memories or confabulations, any of which could be of potential utility or appeal. Delegating certain ANNs within these neural assemblies the task of making value judgments about such confabulations, we form an inventive neural architecture called a “Creativity Machine” (U.S. Patents 5,659,666, 7,454,388, and related U.S. divisional and foreign filings). Within these contemplative computational systems, critic nets govern the statistical placements and magnitudes of disturbances injected into other nets, to induce the highest turnover of potentially useful or meaningful confabulations.
... Oftentimes, this fundamental neural network paradigm takes the form of multiple imagitrons and perceptrons [4,9,10]. Within these neural systems, ideas form as memories and/or confabulations co-activate within distinct imagitrons. ...
Whereas the literature is replete with findings linking creativity and various psychopathologies, the main hypothesis advanced herein is that the neurodynamics of both phenomena are the same. If vindicated, this theory may lead to advanced treatments that could potentially boost creativity as well as safeguard against the associated cognitive and psychological disorders, all through control of just one parameter, the difference between cortical concentrations of excitatory and inhibitory neurotransmitters. Share Link: http://authors.elsevier.com/a/1TSHC15pGbnziG.
... A computational critic, typically taking the form of another neural system, associates each of these emerging patterns with an anticipated pattern-based consequence whose scalar distance, , from some goal pattern is used to modulate the amplitude of the aforesaid synaptic fluctuations. Allowed to run autonomously, such a neural system may eventually arrive at a solution pattern, at which time the metric approaches zero and the synaptic perturbations proportional to it subside (Thaler, 1998). ...
A critical level of synaptic perturbation within a trained, artificial neural system induces the nucleation of novel activation patterns, many of which could qualify as viable ideas or action plans. In building massively parallel connectionist architectures requiring myriad, coupled neural modules driven to ideate in this manner, the need has arisen to shift the attention of computational critics to only those portions of the neural “real estate” generating sufficiently novel activation patterns. The search for a suitable affordance to guide such attention has revealed that the rhythm of pattern generation by synaptically perturbed neural nets is a quantitative indicator of the novelty of their conceptual output, that cadence in turn characterized by a frequency and a corresponding temporal clustering that is discernible through fractal dimension. Anticipating that synaptic fluctuations are tantamount in effect to volume neurotransmitter release within cortex, a novel theory of both cognition and consciousness arises that is reliant upon the rate of transitions within cortical activation topologies.
... The analysis of the electrical properties of PZT ceramics was carried out by a back propagation artificial neural network method (Cai, Xia, Li, & Gui, 2006). The ANN were used in predicting ultra-hard binary compounds (Thaler, 1998), new orthorhombic ABO 3 perovskites (Aleksovska, Dimitrovska, & Kuzmanovski, 2007), the band gap energy and the lattice constant of chalcopyrites (Zeng, Chua, & Wu, 2002), the hexagonal lattice parameters of apatites (Kockan & Evis, 2010). Principal component analysis (PCA) was used in the search for hydrogen storage for AB 5 -alloys (Ye, Xia, Wu, Du, & Zhang, 2002). ...
The review of applications of machine training methods to inorganic chemistry and materials science is presented. The possibility of searching for classification regularities in large arrays of chemical information with the use of precedent-based recognition methods is discussed. The system for computer-assisted design of inorganic compounds, with an integrated complex of databases for the properties of inorganic substances and materials, a subsystem for the analysis of data, based on computer training (including symbolic pattern recognition methods), a knowledge base, a predictions base, and a managing subsystem, has been developed. In many instances, the employment of the developed system makes it possible to predict new inorganic compounds and estimate various properties of those without experimental synthesis. The results of application of this information-analytical system to the computer-assisted design of inorganic compounds promising for the search for new materials for electronics are presented.
... Early on, it became quite evident that the same system, namely, a synaptically perturbed neural network, could provide creative possibilities within any conceptual space. Thus, an associative memory that had been cumulatively exposed to valid chemical formulas produced both novel and plausible theoretical chemistries (Thaler 1998). A perturbed perceptron shown numerous design specifications for consumer products generated plausible new merchandise concepts (Plotkin 2009). ...
Although the definition of the term “creativity” widely varies, recent developments in the field of artificial neural networks (ANN) lend a highly comprehensive model to all accounts of this highly prized cognitive process. From this bottom-up, computational perspective, seminal idea formation results from a noise-driven brain storming session between at least two neural assemblies. In effect, ongoing disturbances both to and within such nets serve to drive a sequence of activation patterns in a process tantamount to stream of consciousness. At sufficiently intense disturbance levels, memories and their interrelationships degrade into false memories or confabulations, any of which could be of potential utility or appeal. If another ANN is provided to make this value judgment, we form an inventive neural architecture called a “Creativity Machine” (U.S. Patents 5,659,666, 7,454,388, and related U.S. divisional and foreign filings). Within such contemplative computational systems, the latter network may be allowed governance over the statistical placement and magnitudes of such disturbances, so as to induce the highest turnover of potentially useful or meaningful confabulations.
According to this simple, elegant, and working model, creativity may be attributed to the failure of biological neural networks to reconstruct memories of direct experience when exposed to nature’s ubiquitous disordering effects, as other “wetware” opportunistically exploits such mistakes and pragmatically perfects the underlying network flaws.
... The overall result is that the network fails to activate into its learned output exemplars at its terminal layer. In effect, the net is then generating false memories or confabulations as its synaptic connections are continually perturbed (Thaler 1998). ...
In Alan Turing’s landmark paper, “Computing Machinery and Intelligence,” the famous cyberneticist takes the position that machines will inevitably think, supplied adequate storage, processor speed, and an appropriate program. Herein we propose the solution to the latter prerequisite for contemplative machine intelligence, the required algorithm, illustrating how it weathers the criticism well anticipated by Turing that a computational system can never attain consciousness.
A novel form of neurocomputing allows machines to generate new concepts along with their anticipated consequences, all encoded as chained associative memories. Knowledge is accumulated by the system through direct experience as network chaining topologies form in response to various environmental input patterns. Thereafter, random disturbances to the connections joining these nets promote the formation of alternative chaining topologies representing novel concepts. The resulting ideational chains are then reinforced or weakened as they incorporate nets containing memories of impactful events or things. Such encodings of entities, actions, and relationships as geometric forms composed of artificial neural nets may well suggest how the human brain summarizes and appraises the states of nearly a hundred billion cortical neurons. It may also be the paradigm that allows the scaling of synthetic neural systems to brain-like proportions to achieve sentient artificial general intelligence (SAGI).
The purpose of this paper is to describe a generalized approach to developing computational tools capable of: fusing data (soft and hard) obtained from in service systems, bench/lab scale experiments and multi-phase multi-scale impact dynamics simulations of in situ nano composite materials; accurately representing the response surface(s) associated with the large number of coupled parameters that govern the impact dynamics of; autonomously employ techniques such as Design of Experiments and Response Surface Methodology to design and execute virtual experiments in order to adaptively optimize material performance. In addition, the potential for solving inverse problems associated with material phase optimization, micro-structural optimization and adaptive response characterization as it is related to hypervelocity impact of nano- composite materials will be discussed.
The problems of predicting not yet synthesized intermetallic compounds are discussed. It is noted that the use of classical physicochemical analysis in the study of multicomponent metallic systems is faced with the complexity of presenting multidimensional phase diagrams. One way of predicting new intermetallics with specified properties is the use of modern processing technology with application of teaching of image recognition by the computer. The algorithms used most often in these methods are briefly considered and the efficiency of their use for predicting new compounds is demonstrated.
An artificial neural network cascade, containing 16 individual network modules and approximately 1,000 processing units, has produced an interactive database of nearly a quarter million potential binary and ternary chemical systems. While many of these hypothetical materials are anticipated to be thermodynamically stable, they are most likely kinetically inaccessible via typical bulk chemistry routes. However, since modem thin film technology allows a wide range of exotic compositions and stoichiometries via deposition, surface treatments, and nano-fabrication, it is anticipated that this newly acquired theoretical database will form a comprehensive road map to the formation of previously unattainable materials that offer significant technological advantages. Further, with the suite of available coating materials greatly expanded, thin film designers now have at their disposal the means to implement multilayer and composite thin film device designs that fulfill a much broader range of performance requirements and that are ideally matched to both underlying substrate and external working environment.
In this review, we highlight recent applications of machine learning to virtual screening, focusing on the use of supervised techniques to train statistical learning algorithms to prioritize databases of molecules as active against a particular protein target. Both ligand-based similarity searching and structure-based docking have benefited from machine learning algorithms, including naïve Bayesian classifiers, support vector machines, neural networks, and decision trees, as well as more traditional regression techniques. Effective application of these methodologies requires an appreciation of data preparation, validation, optimization, and search methodologies, and we also survey developments in these areas.
A synaptically perturbed neural network forms an efficient search engine within and around any conceptual space upon which it has been trained. By monitoring the temporal distribution of concepts emerging from such a system, we discover a quantitative agreement with the measured rhythm of human cognition, creative or otherwise. Closer examination of this transparent connectionist search engine suggests that much of human creativity may be attributed to the failure of cortical networks to activate into known memories as these networks perform vector completion upon their own internal disturbances. In lieu of intact memory activation, the networks produce a stream of degraded memories, now constituting what we commonly refer to as "ideas," that are filtered for utility and interest by attendant cortical networks.
Recent experiments have shown that a trained chaotic network supervised by an associative net may produce human level discovery and invention. Extending this so-called “Creativity Machine Paradigm” to more complex cascades of chaotic networks not only allows for more ambitious forms of discovery such as juxtapositional invention, but also suggests a nomenclature and a convenient classification scheme for creative human achievements.
Recent experiments have shown that a trained chaotic network supervised by an associative net may produce human level discovery and invention. Extending this so-called “Creativity Machine Paradigm” to more complex cascades of chaotic networks not only allows for more ambitious forms of discovery such as juxtapositional invention, but also suggests a nomenclature and a convenient classification scheme for creative human achievements.
We present a first-principles pseudopotential study of the structural and electronic properties of β-Si3N4 and the hypothetical compound β-C3N4. β-C3N4, which is C3N4 in the β-Si3N4 structure, with C substituted for Si, is used as a prototype for investigating the properties of possible covalent C-N solids. The calculated lattice constant, bulk modulus, and electronic band structure of β-Si3N4 are in good agreement with experimental results. This gives support for the predicted properties of β-C3N4. The bulk modulus of β-C3N4 is found to be comparable to diamond, and its moderately large cohesive energy suggests that the prototype structure may be metastable. Although the crystal structure and the valencies of the constituent atoms are similar in β-Si3N4 and β-C3N4, the electronic bonding properties in these two solids are found to differ. The large core size and repulsive p pseudopotential of the second-row element, Si, results in a more ionic Si-N bond compared with a covalent C-N bond.
Many of the most interesting developments in semiconductor physics that have occurred in the last few years and that are anticipated
in the next few years appear to lie in the realm between physics and chemistry. In this article I have tried to show how this
realm can be treated accurately and realistically within the framework of theory.
We describe a hypothetical new phase of carbon. The proposed structure is a rigid three-dimensional network with sp2 (threefold) coordination only. Tight-binding calculations indicate that the proposed material is a metastable metal significantly harder than diamond. Although it is unclear at present whether such a material can be produced in significant quantity, we speculate that it (and related structures) plays a role in the chemical vapor deposition of diamond films.
A simple pattern association network, trained to convert any of eight three-bit input patterns to corresponding 3 × 3 pixel patterns, is destroyed by the random pruning of its connection weights. Within such “deaths” we see the frequent appearance of the trained output patterns independent of the application of the corresponding inputs. Such events, which we shall call “virtual inputs, ” increase infrequency when input layer weights are pruned in favor of those in the output layer. We ultimately attribute the virtual inputs to a neural network completion process in which the pattern of zeros produced by the stochastic decay are interpreted by the largely intact hidden and output layers as the application of any of the eight original training input vectors to the input units. After isolating the essential mechanisms producing virtual inputs, we generalize this phenomenon to a wide range ofparallel distributed systems.
An empirical model and an ab initio calculation of the bulk moduli for covalent solids are used to suggest possible new hard materials. The empirical model indicates that hypothetical covalent solids formed between carbon and nitrogen are good candidates for extreme hardness. A prototype system is chosen and a first principles pseudopotential total energy calculation on the system is performed. The results are consistent with the empirical model and show that materials like the prototype can have bulk moduli comparable to or greater than diamond. It may be possible to synthesize such materials in the laboratory.
A Field Guide to Rocks and Minerals, The Riverside intermetallics generally segregate into separate hardness Press, Cambridge, 1960. families, each represented by diagonal trend lines
- F H Pough
F.H. Pough, A Field Guide to Rocks and Minerals, The Riverside intermetallics generally segregate into separate hardness Press, Cambridge, 1960. families, each represented by diagonal trend lines
) (1995) 592. now possessing a residual constant term reflecting the
- S L Thaler
- Bull
- Am
S.L. Thaler, Bull. Am. Phys. Soc. 40(1) (1995) 592. now possessing a residual constant term reflecting the
Neural networks that create and discover, PC AI, hardness of the underlying homonuclear host lattice May
- S L Thaler
S.L. Thaler, Neural networks that create and discover, PC AI, hardness of the underlying homonuclear host lattice May / June, 1996, pp. 16-21.
The spreadsheet implemented `Creativity Machine' - its construction, function, and current successes
- S L Thaler