Deep Learning Accelerators' Configuration Space Exploration Effect on Performance and Resource Utilization: A Gemmini Case Study.

Though custom deep learning (DL) hardware accelerators are attractive for making inferences in edge computing devices, their design and implementation remain a challenge. Open-source frameworks exist for exploring DL hardware accelerators. Gemmini is an open-source systolic array generator for agile DL accelerator exploration. This paper details the hardware/software components generated using Gemmini. The general matrix-to-matrix multiplication (GEMM) of different dataflow options, including output/weight stationary (OS/WS), was explored in Gemmini to estimate the performance relative to a CPU implementation. The Gemmini hardware was implemented on an FPGA device to explore the effect of several accelerator parameters, including array size, memory capacity, and the CPU/hardware image-to-column (im2col) module, on metrics such as the area, frequency, and power. This work revealed that regarding the performance, the WS dataflow offered a speedup of 3× relative to the OS dataflow, and the hardware im2col operation offered a speedup of 1.1× relative to the operation on the CPU. For hardware resources, an increase in the array size by a factor of 2 led to an increase in both the area and power by a factor of 3.3, and the im2col module led to an increase in area and power by factors of 1.01 and 1.06, respectively.

Lightweight and Energy-Efficient Deep Learning Accelerator for Real-Time Object Detection on Edge Devices.

Tiny machine learning (TinyML) has become an emerging field according to the rapid growth in the area of the internet of things (IoT). However, most deep learning algorithms are too complex, require a lot of memory to store data, and consume an enormous amount of energy for calculation/data movement; therefore, the algorithms are not suitable for IoT devices such as various sensors and imaging systems. Furthermore, typical hardware accelerators cannot be embedded in these resource-constrained edge devices, and they are difficult to drive real-time inference processing as well. To perform the real-time processing on these battery-operated devices, deep learning models should be compact and hardware-optimized, and hardware accelerator designs also have to be lightweight and consume extremely low energy. Therefore, we present an optimized network model through model simplification and compression for the hardware to be implemented, and propose a hardware architecture for a lightweight and energy-efficient deep learning accelerator. The experimental results demonstrate that our optimized model successfully performs object detection, and the proposed hardware design achieves 1.25× and 4.27× smaller logic and BRAM size, respectively, and its energy consumption is approximately 10.37× lower than previous similar works with 43.95 fps as a real-time process under an operating frequency of 100 MHz on a Xilinx ZC702 FPGA.

Religion and Misconduct Among Prison Inmates in South Korea.

Although faith-based programs are present in most prisons for offender rehabilitation, the effect of religion on prison inmates remains an understudied topic. In addition, existing research shows mixed results about the religious effect. The present study intends to not only advance the understanding of inmate's prison misconduct but also examine whether religion is likely to contribute to reducing the risk of misconduct using a non-Western sample of inmates. To assess the relationship between inmates' religion and prison misconduct, we applied negative binomial regression to analyze survey data from 986 Korean adult male inmates. Results showed that inmates who had a religious affiliation with Catholicism or Buddhism were less likely to report prison misconduct than those who had no religion. In addition, the inverse relationship was observed whether inmates had participated in religion before incarceration or came to participate in religion while incarcerated, depending on religious denomination.

Religion and Responsibility-Taking Among Offenders in Colombia and South Africa: A Qualitative Assessment of a Faith-Based Program in Prison.

This paper examines whether religion contributes to offenders taking responsibility for crimes. Specifically, we assessed whether participation in The Prisoner's Journey (TPJ), a bible study program, increased or decreased responsibility-taking. We also examined whether religious offenders that did not participate in TPJ were likely to take responsibility for their offenses. For this study, we conducted a quasi-experiment in two Colombian and five South African prisons from 2018 to 2019, collecting data from personal interviews with a total of 73 inmates-42 TPJ participants and 31 non-participants-before and after the program. Offenders frequently offered subtle accounts of responsibility that incorporated their own agency with other factors. Highly religious offenders were equally likely to take responsibility, and in some cases participation in TPJ heightened responsibility. In sum, this paper presents evidence that religious beliefs and practice are commensurate with responsibility-taking and desistance from crime.

Religion and Rehabilitation as Moral Reform: Conceptualization and Preliminary Evidence.

We examine how religion contributes to rehabilitation, which we conceptualize as moral reform and operationalize in terms of self-identity, existential belief, and character. We hypothesize that religion contributes to identity transformation, a sense of meaning and purpose in life, and virtue development. We also hypothesize that faith-based rehabilitation reduces negative emotions and the risk of interpersonal aggression. We conducted a quasi-experiment on a faith-based program in a state jail and a maximum-security prison in Texas, using a convenience sample of male inmates. To test our hypotheses, we compare inmates who graduated the program with those who did not and applied manifest-variable structural equation modeling to analyze data from pretest and posttest surveys. Program participation was linked to an increase in religiosity, which contributed to identity transformation (cognitive and emotional transformations and crystallization of discontent), the perceived presence of meaning and purpose in life, and virtues (including self-control, compassion, and forgiveness). Faith-based rehabilitation in turn reduced state depression and anxiety and the probability of engaging in aggression toward another inmate. This study provides preliminary evidence of religion's rehabilitative effect on offenders; findings which hold promise for prison administrators looking for creative ways to support evidence-based and cost-effective approaches to rehabilitation within the correctional system. Supplementary Information: The online version contains supplementary material available at 10.1007/s12103-022-09707-3.

Assessing a Faith-Based Program for Trauma Healing Among Jail Inmates: A Quasi-Experimental Study.

This paper assesses a faith-based, short-term program for trauma healing among incarcerated individuals, "Correctional Trauma Healing Program" (CTHP). We hypothesized that participation in the CTHP would reduce negative consequences of lifetime trauma: symptoms of PTSD, state depression, state anger, suicidal ideation, and the risk of interpersonal aggression. We also hypothesized that the reduction, if found, would be partly attributable to anticipated program outcomes (a decrease in vengefulness and an increase in religiosity, forgiveness, perceived forgiveness of God, gratitude to God, and perceived positive impact of the Bible). To test our hypotheses, we conducted a quasi-experimental study of 349 jail inmates in Virginia. Manifest-variable structural equation modeling was applied to analyze data from pretest and posttest surveys. As hypothesized, the CTHP reduced the negative consequences of trauma by increasing religiosity and other positive attributes and decreasing vengefulness directly and/or indirectly via increased religiosity. Implications and limitations of our study are discussed.

Perceptions of Accountability to God and Psychological Well-Being Among US Adults.

This study examines whether accountability to God is positively associated with four measures of psychological well-being-happiness, mattering to others, dignity, and meaning-among US adults. It also tests the possibility that prayer moderates these associations. Data from the 2017 Values and Beliefs of the American Public Survey (n = 1251) were analyzed using multivariate regression. Findings provided support for an association between accountability to God and mattering to others, dignity, and meaning in fully controlled models, and for happiness when religious controls were excluded. They also showed that these relationships were stronger among those who prayed frequently compared with those who did not. Overall, these findings shed light on a new concept-accountability to God-including its association with psychological well-being.

Noise-Aware and Light-Weight VLSI Design of Bilateral Filter for Robust and Fast Image Denoising in Mobile Systems.

The range kernel of bilateral filter degrades image quality unintentionally in real environments because the pixel intensity varies randomly due to the noise that is generated in image sensors. Furthermore, the range kernel increases the complexity due to the comparisons with neighboring pixels and the multiplications with the corresponding weights. In this paper, we propose a noise-aware range kernel, which estimates noise using an intensity difference-based image noise model and dynamically adjusts weights according to the estimated noise, in order to alleviate the quality degradation of bilateral filters by noise. In addition, to significantly reduce the complexity, an approximation scheme is introduced, which converts the proposed noise-aware range kernel into a binary kernel while using the statistical hypothesis test method. Finally, blue a fully parallelized and pipelined very-large-scale integration (VLSI) architecture of a noise-aware bilateral filter (NABF) that is based on the proposed binary range kernel is presented, which was successfully implemented in field-programmable gate array (FPGA). The experimental results show that the proposed NABF is more robust to noise than the conventional bilateral filter under various noise conditions. Furthermore, the proposed VLSI design of the NABF achieves 10.5 and 95.7 times higher throughput and uses 63.6-97.5% less internal memory than state-of-the-art bilateral filter designs.

Prisoners Helping Prisoners Change: A Study of Inmate Field Ministers Within Texas Prisons.

Research on incarcerated offenders trained to help prisoners change is rare because programs that equip inmates with practical capacities for helping others rehabilitate in prison hardly exist. An exception is the Field Ministry program in Texas, which enlists inmates who have graduated from a prison-based seminary to work as "Field Ministers" and serve other inmates in various capacities. We hypothesize that inmate exposure to Field Ministers is inversely related to antisocial factors and positively to prosocial ones. We applied manifest-variable structural equation modeling to analyze data from a survey of a random sample of male inmates at three maximum-security prisons where the Field Ministry program operated. We found that inmates exposed more frequently to the Field Ministry and for a longer time period tended to report lower levels of criminological risk factors and aggressiveness and higher levels of virtues and predictors of human agency as well as religiosity and spirituality.

Religiosity, Crime, and Drug Use Among Juvenile Offenders: A Test of Reciprocal Relationships Over Time.

This article examines whether an individual's religiosity has reciprocal relationships with crime and drug use among juvenile offenders. Structural equation modeling is applied to analyze 11-wave panel data from a study of juveniles adjudicated or found guilty of a serious offense in two states. Offenders' religiosity is measured both objectively (participation in religious activities) and subjectively (religious salience, experiences, and efficacy). While holding constant an offender's exposure time (the proportion of time on the street), previous levels of crime and drug use, and sociodemographic controls, this study found the relationship between religiosity and crime (i.e., nondrug offending) to be either bidirectional or unidirectional. The relationship between religiosity and drug use (binge drinking, marijuana use, and hard drug use) is, however, unidirectional over time. When unidirectional relationship is found, it is religiosity that decreases crime and drug use, not the other way around. Implications of findings are discussed.

"First Stop Dying".

This article offers an ethnographic account of the "self-projects" of inmate graduates of Louisiana State Penitentiary's (aka "Angola's") unique prison seminary program. Angola's Inmate Minister program deploys seminary graduates in bivocational pastoral service roles throughout America's largest maximum-security prison. Drawing upon the unique history of Angola, inmates establish their own churches and serve in lay-ministry capacities in hospice, cellblock visitation, tier ministry, officiating inmate funerals, and through tithing with "care packages" for indigent prisoners. Four themes of positive criminology prominently emerge from inmate narratives: (a) the importance of respectful treatment of inmates by correctional administrations, (b) the value of building trusting relationships for prosocial modeling and improved self-perception,

Does Alcohol or Delinquency Help Adolescents Feel Better Over Time? A Study on the Influence of Heavy Drinking and Violent/Property Offending on Negative Emotions.

Conceptualizing adolescent drinking and delinquency as adaptations to strain, we explore whether they (a) decrease or increase the probability of feeling depression and anxiety later and (b) ameliorate or aggravate the effect of strain on the negative emotions over time. These relationships are also examined for gender differences by analyzing data separately for males and females as well as both combined. We conducted ordinary least squares regression analysis of panel data from the first two waves of the National Longitudinal Study of Adolescent to Adult Health. Heavy drinking and serious delinquency were found to increase the probability of feeling depression and anxiety later, whereas they tend to ameliorate the emotionally deleterious effect of strain for males and, to a lesser extent, females.

A General Strain Approach to Psychological and Physical Bullying: A Study of Interpersonal Aggression at School.

Although studies have been conducted to examine the applicability of Agnew's general strain theory (GST) to the explanation of school bullying, GST research on the phenomenon remains limited in number and scope. To fill this gap in research, using data from a sample of 296 middle school students in a southwestern state of the United States, this article examined whether different types of strain and negative emotions are positively associated with psychological, physical, and general bullying. Overall findings of negative binomial regression analyses tended to be consistent with our expectations, while some aspects of GST received more empirical support than others. Strains and negative emotions were mostly related positively either to psychological or physical bullying, with negative emotions, anger and depression, partly mediating the strain-bullying relationship. However, we found mixed patterns of interactions among strains, negative emotions, and conditioning factors. Theoretical and practical implications were discussed.

A fully integrated sensor SoC with digital calibration hardware and wireless transceiver at 2.4 GHz.

A single-chip sensor system-on-a-chip (SoC) that implements radio for 2.4 GHz, complete digital baseband physical layer (PHY), 10-bit sigma-delta analog-to-digital converter and dedicated sensor calibration hardware for industrial sensing systems has been proposed and integrated in a 0.18-µm CMOS technology. The transceiver's building block includes a low-noise amplifier, mixer, channel filter, receiver signal-strength indicator, frequency synthesizer, voltage-controlled oscillator, and power amplifier. In addition, the digital building block consists of offset quadrature phase-shift keying (OQPSK) modulation, demodulation, carrier frequency offset compensation, auto-gain control, digital MAC function, sensor calibration hardware and embedded 8-bit microcontroller. The digital MAC function supports cyclic redundancy check (CRC), inter-symbol timing check, MAC frame control, and automatic retransmission. The embedded sensor signal processing block consists of calibration coefficient calculator, sensing data calibration mapper and sigma-delta analog-to-digital converter with digital decimation filter. The sensitivity of the overall receiver and the error vector magnitude (EVM) of the overall transmitter are -99 dBm and 18.14%, respectively. The proposed calibration scheme has a reduction of errors by about 45.4% compared with the improved progressive polynomial calibration (PPC) method and the maximum current consumption of the SoC is 16 mA.