Artificial Intelligence Processing Research Articles

APPLICATION OF METHODS OF THE ARTIFICIAL INTELLIGENCE BY PROCESS PLANNING

The results of the analysis of the current state of the level of automation of process planning at the enterprises of the machine-building industry are presented in the paper. A comparative analysis of two methods of organizing the work of the technological service of the enterprise is carried out, the analysis is carried out on the example of the studied standard product. An enlarged method of calculating the time spent on process planning are presented. The calculations of reducing labor intensity, expressed in monetary terms, are given. The plan of introduction of artificial intelligence in process planning at the enterprises of machine-building industry are presented. The possibilities of further growth of production automation based on computer-aided design systems and artificial intelligence technologies are considered. The analysis of each stage of process planning allows us to conclude that the digitalization of production will allow us to move to a new level of automation, combining individual automation elements into a single integrated system. In conclusion, performance indicators from the introduction of artificial intelligence technologies, increased productivity and reduced labor intensity are noted.

A Multi-Core Controller for an Embedded AI System Supporting Parallel Recognition.

Recent advances in artificial intelligence (AI) technology encourage the adoption of AI systems for various applications. In most deployments, AI-based computing systems adopt the architecture in which the central server processes most of the data. This characteristic makes the system use a high amount of network bandwidth and can cause security issues. In order to overcome these issues, a new AI model called federated learning was presented. Federated learning adopts an architecture in which the clients take care of data training and transmit only the trained result to the central server. As the data training from the client abstracts and reduces the original data, the system operates with reduced network resources and reinforced data security. A system with federated learning supports a variety of client systems. To build an AI system with resource-limited client systems, composing the client system with multiple embedded AI processors is valid. For realizing the system with this architecture, introducing a controller to arbitrate and utilize the AI processors becomes a stringent requirement. In this paper, we propose an embedded AI system for federated learning that can be composed flexibly with the AI core depending on the application. In order to realize the proposed system, we designed a controller for multiple AI cores and implemented it on a field-programmable gate array (FPGA). The operation of the designed controller was verified through image and speech applications, and the performance was verified through a simulator.

Hunch Mining

Intuition and hunches are important tools for experts who make time-critical highly complex decisions in turbulent environments. However, hunches are also elusive and exist below the surface when not being used for immediate decision making. These latent hunches are valuable and can lead to creative solutions outside of a crisis. This paper uses a demonstration experiment to show how hunches can be “mined” using revealed causal mapping techniques. Fifty chief information officers and 88 programmer/analysts were interviewed during a very turbulent time in their organizations. The hunch mining results indicate that hunches can be found throughout experts' communications and can serve as valuable insights into strategic and tactical decision making. This research serves as the first manual step in developing natural language processing artificial intelligence tools for automated hunch mining in expert communications.

Editorial Comment: On the Possibilities of AI, Performance Studies, and Digital Cohabitation

Editorial Comment:On the Possibilities of AI, Performance Studies, and Digital Cohabitation Sean Metzger Over the three years that I initially conceived and eventually edited this special issue, I have become convinced of the pressing need to think through the intersections of performance studies and artificial intelligence (AI). Many factors shape this belief, but they might be summarized in a relatively succinct formulation: Code is performative, and interfaces are theatrical; the elaborations of these claims reveal the shifting of both human and machine ontologies as the networks we inhabit proliferate. To a large extent, the constituent essays in this volume complicate and extend such ideas. The following editorial grounds these elaborations and provides one pathway for viewing their collective charge. Scholars of digital media have established the performativity of computer code.1 For example, Lisa Nakamura has pointed to the performative acts that occur as users create avatars in virtual environments.2 N. Katherine Hayles has explored performativity to account for how computational theory might shift the linguistic framework of this concept provided by J. L. Austin: Computational theory treats computer languages as if they were, in Austin's terms, performative utterances. Although material changes do take place when computers process code (magnetic polarities are changed on a disk), it is the act of attaching significance to these physical changes that constitutes computation as such. … Whereas in performative utterances saying is doing because the action performed is symbolic in nature and does not require physical action in the world, at the basic level of computation doing is saying because physical actions also have a symbolic dimension that corresponds directly to computation.3 From a related perspective, Wendy Chun refers to code as "an inhumanly perfect 'performative' uttered by no one. Unlike any other law or performative utterance, code almost always does what it says because it needs no human acknowledgment."4 Code within this context generally denotes steps or commands for a computer to execute. Coding pertains to, but is not limited to, AI, which relies on sophisticated algorithms. A high-performance algorithm generally refers to any complex formula that "optimally organizes the intake of input and the production of output."5 AI's processes of decision-making often remain unknown, a black box. The ways such algorithms [End Page xiii] increasingly structure our lives have become ever more vexing, leading to what Safiya Noble has called "algorithmic oppression" as "fundamental to the operating system of the web" in her writing on search engines.6 In Noble's research, the results of algorithmic search processes yielded disturbing presentations of bodies for consumption; put otherwise, the histrionic display of Black women on porn sites substituted one vision of femininity for the one that Noble sought (specifically, material that might be interesting to some of the youth in her family). Typing "black girls" in a search engine in 2010 produced unexpected iterations of ostensibly Black womanhood. This example demonstrates both the performative nature of the algorithm and also points to the theatricality of the interface. Broadly speaking, discussions about computers often use "screen" as shorthand for the specific structure called the graphic user interface (GUI). However, as Johanna Drucker has explained, such usage eschews the complexities of human–computer interaction (HCI) that involve keyboards and the like used for "remediating common tasks as formal instructions coded into a plan of menus or buttons."7 Automated interfaces involve aural, visual, and tactile sensory modalities.8 As technology advanced, "graphical icons … no longer just looked like objects; they could also mimic the behavior of the things they resembled."9 Many implications follow from this statement, but Drucker offers one illustration relevant to theatricality in particular: "The GUI is a mediating scrim, a boundary space in which we interact with an abstraction of computation … neither the file folders nor the other objects in the screen space are things; they are icons that represent behaviors and actions we want to perform … and they interpellate a user through disciplinary and scopic regimes."10 She further explains that the GUI effects a "constitutive exchange" that involves "cognitive adaptation and change" through which the interface facilitates "individual and collective subject formation."11 Drucker does not mention theatre studies or...

Special issue on SoC and AI processors

Special issue on SoC and AI processors

Intellino: Processor for Embedded Artificial Intelligence

The development of computation technology and artificial intelligence (AI) field brings about AI to be applied to various system. In addition, the research on hardware-based AI processors leads to the minimization of AI devices. By adapting the AI device to the edge of internet of things (IoT), the system can perform AI operation promptly on the edge and reduce the workload of the system core. As the edge is influenced by the characteristics of the embedded system, implementing hardware which operates with low power in restricted resources on a processor is necessary. In this paper, we propose the intellino, a processor for embedded artificial intelligence. Intellino ensures low power operation based on optimized AI algorithms and reduces the workload of the system core through the hardware implementation of a neural network. In addition, intellino’s dedicated protocol helps the embedded system to enhance the performance. We measure intellino performance, achieving over 95% accuracy, and verify our proposal with an field programmable gate array (FPGA) prototyping.

Towards silicon photonic neural networks for artificial intelligence

Brain-inspired photonic neural networks for artificial intelligence have attracted renewed interest. For many computational tasks, such as image recognition, speech processing and deep learning, photonic neural networks have the potential to increase the computing speed and energy efficiency on the orders of magnitude compared with digital electronics. Silicon Photonics, which combines the advantages of electronics and photonics, brings hope for the large-scale photonic neural network integration. This paper walks through the basic concept of artificial neural networks and focuses on the key devices which construct the silicon photonic neuromorphic systems. We review some recent important progress in silicon photonic neural networks, which include multilayer artificial neural networks and brain-like neuromorphic systems, for artificial intelligence. A prototype of silicon photonic artificial intelligence processor for ultra-fast neural network computing is also proposed. We hope this paper gives a detailed overview and a deeper understanding of this emerging field.

A time-domain analog weighted-sum calculation circuit using ferroelectric-gate field-effect transistors for artificial intelligence processors

Time-domain analog computing with transient states (TACT) approaches have been proposed to realize high performance artificial intelligence (AI) processors, which have over 100 times higher energy efficiency than the latest digital AI processors. Such approaches require analog memory devices that can hold and control very high resistance on the order of giga-ohms to tera-ohms, with nonvolatility and rectification. Memory devices with a field-effect transistor (FET) structure can operate in a subthreshold region to realize such high resistance, and can also rectify the current using FET operation. In this paper, we propose using ferroelectric-gate FETs (FeFETs) as a candidate for such devices. We design crossbar circuit architectures for using them, and show measurement results for device characteristics as analog memory devices with pulse control based on TACT approaches, and evaluation results for weighted-sum calculation operation in proof-of-concept FeFET circuits.

Flip-Chip Flux Evolution

Abstract Flip-chip assembly accounts for more than 80% of the advanced packaging technology platform, compared to fan-in, fan-out, embedded die, and through silicon via (TSV). Flip-chip interconnect remains a critical assembly process for large die used in artificial intelligence processors; thin die that warps at elevated temperatures; heterogeneous integration in SiP applications; flip-chip on leadframe; and MicroLED die usage. This paper will first outline trends in evolving flip-chip and direct chip placement (DCP) technology, then will examine the changing nature of the solder bump, the interconnect itself, and the substrate. Many variables of the flip-chip assembly process will be discussed, including standard solder bumps to micro Cu-pillar bumps with different alloys; different pad surface finishes of Cu OSP, NiAu, and solder on pad (SOP); and from regular pads on substrates to bond-on-trace applications. A major focus will be on flip-chip assembly methods, from old C4 conventional reflow processing to thermocompression bonding (TCB), and the latest laser assisted bonding (LAB) technology, with an emphasis on how the usage of different technologies necessitates different assembly materials, especially fluxes. Flip-chip fluxes such as the commonly used water-washable flux, the standard no-clean flux, and the ultra-low residue flux, and how these fluxes react to different processing methods, will be an area of discussion. Finally, the paper will examine the need for increased reliability as the technology inevitably moves into the high-volume, zero-defect arena of automotive electronics.

A 0.55 V 1.1 mW Artificial Intelligence Processor With On-Chip PVT Compensation for Autonomous Mobile Robots

Autonomous mobile robots are receiving a lot of attention for many applications, such as package delivery and smart surveillance, however, the battery capacity is limited to implement intelligent decision making in robots because of the heavy computational costs. In this paper, an ultra-low-power artificial intelligence processor (AIP) is proposed for real-time decision making of autonomous mobile robots. To achieve low power consumption while maintaining high performance, it adopts four key features: 1) an 8-thread tree search processor for real-time path planning; 2) a reinforcement learning accelerator for the avoidance of unexpected obstacles; 3) a 3-level transposition table cache for the reduction of duplicated computations; and 4) a PVT compensation circuit for the stable operation at near-threshold voltage. The proposed 16 mm $^{{{2}}}$ AIP is fabricated using 65-nm triple-well CMOS technology. It consumes only 1.1 mW at 0.55 V supply voltage and 7 MHz operating frequency, and 151 mW at 1.2 V supply voltage and 245 MHz operating frequency. The AIP achieves fast search speed (470 000 search/s) and low energy consumption (79 nJ/search), and it is successfully applied to a battery-powered robot system for autonomous navigation without any collision in dynamic indoor environments.

Enhanced Clustering Techniques for Hyper Network Planning using Minimum Spanning Trees and Ant-Colony Algorithm

Problem statement: The process of network planning is divided into two sub steps. The first step is determining the location of the Multi Service Access Node (MSAN). The second step is the construction of subscriber network lines from MSAN to subscribers to satisfy optimization criteria and design constraints. Due to the complexity of this process artificial intelligence and clustering techniques have been successfully deployed to solve many problems. The problems of the locations of MSAN, the cabling layout and the computation of optimum cable network layouts have been addressed in this study. The proposed algorithm, Clustering density-Based Spatial of Applications with Noise original, minimal Spanning tree and modified Ant-Colony-Based algorithm (CBSCAN-SPANT), used two clustering algorithms which are density-based and agglomerative clustering algorithm using distances which are shortest paths distance and satisfying the network constraints. This algorithm used wire and wireless technology to serve the subscribers demand and place the switches in a real optimal place. Approach: The density-based Spatial Clustering of Applications with Noise original (DBSCAN) algorithm has been modified and a new algorithm (NetPlan algorithm) has been proposed by the author in a recent work to solve the first step in the problem of network planning. In the present study, the NetPlan algorithm is modified by introduce the modified Ant-Colony-Based algorithm to find the optimal path between any node and the corresponding MSAN node in the first step of network planning process to determine nodes belonging to each cluster. The second step, in the process of network planning, is also introduced in the present study. For each cluster, the optimal cabling layout from each MSAN to the subscriber premises is determining by introduce the Prime algorithm which construct minimal spanning tree. Results: Experimental results and analysis indicate that the CBSCAN-SP-ANT algorithm was effective, leads to minimum costs for network construction and make the best grade of service. Conclusion: Using mobile network to serve the area with low density is decreasing the cost of design the fixed wire network. Also, using the modified ANT algorithm and minimum spanning tree, are helping to construct the cable layout from each MSAN to subscribers when the network is complicated and the number of intersections and streets are large.

Artificial intelligence in process engineering—current state and future trends

Recent advances in artificial intelligence have changed the fundamental assumptions upon which the progress of computer-aided process engineering (modeling and methodologies) during the last 30 yr has been founded. Thus, in certain instances, numerical computations today constitute inferior alternatives to qualitative and/or semi-quantitative models and procedures which can capture and utilize more broadly- based sources of knowledge. In this paper it will be shown how process development and design, as well as planning, scheduling, monitoring, analysis and control of process operations can benefit from improved knowledge-representation schemes and advanced reasoning control strategies. It will also be argued that the central challenge coming from research advances in artificial intelligence is “modeling the knowledge”, i.e. modeling: (a) physical phenomena and the systems in which they occur; (b) information handling and processing systems; and (c) problem-solving strategies in design, operations and control. Thus, different strategies require different forms of declarative knowledge, and the success or failure of various design, planning, diagnostic and control systems depends on the extent of actively utilizable knowledge. Furthermore, this paper will outline the theoretical scope of important contributions from AI and what their impact has been and will be on the formulation and solution of process engineering problems.

Artificial intelligence in process planning

Both the changing demands of the market and keen competition among manufacturers call for a more flexible approach to manufacturing. Adopting conventional computing methods is not adequate to meet the sophisticated needs of the modern manufacturing industry. This paper describes the approach of applying artificial intelligence techniques to process planning. An analogy between artificial intelligence planning and process planning is made. To illustrate this, the ‘blocks world’ is compared with the assembly of the piston connecting rod of car engines.