检索结果-内蒙古大学图书馆

Consciousness in Artificial Intelligence: Insights from the science of Consciousness

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Butlin, Patrick Long, Robert Elmoznino, Eric Bengio, Yoshua Birch, Jonathan Constant, Axel Deane, George Fleming, Stephen M. Frith, Chris Ji, Xu Kanai, Ryota Klein, Colin Lindsay, Grace Michel, Matthias Mudrik, Liad Peters, Megan A.K. Schwitzgebel, Eric Simon, Jonathan VanRullen, Rufin Future of Humanity Institute University of Oxford United Kingdom Center for AI Safety United States University of Montreal MILA - Quebec AI Institute Canada Centre for Philosophy of Natural and Social Science London School of Economics and Political Science United Kingdom School of Engineering and Informatics The University of Sussex Centre de Recherche en Éthique University of Montreal United Kingdom Department of Philosophy University of Montreal Canada Department of Experimental Psychology Wellcome Centre for Human Neuroimaging University College London United Kingdom Wellcome Centre for Human Neuroimaging University College London Institute of Philosophy University of London United Kingdom Araya Inc. School of Philosophy The Australian National University Australia Department of Psychology Center for Data Science New York University United States Center for Mind Brain and Consciousness New York University United States School of Psychological Sciences and Sagol School of Neuroscience Tel-Aviv University CIFAR Program in Brain Mind and Consciousness Israel Department of Cognitive Sciences University of California Irvine and CIFAR Program in Brain Mind and Consciousness United States Department of Philosophy University of California Riverside United States Centre de Recherche Cerveau et Cognition CNRS Université de Toulouse France

Whether current or near-term AI systems could be conscious is a topic of scientific interest and increasing public concern. This report argues for, and exemplifies, a rigorous and empirically grounded approach to AI consciousness: assessing existing AI systems in detail, in light of our best-supported neuroscientific theories of consciousness. We survey several prominent scientific theories of consciousness, including recurrent processing theory, global workspace theory, higher-order theories, predictive processing, and attention schema theory. From these theories we derive"indicator properties" of consciousness, elucidated in computational terms that allow us to assess AI systems for these properties. We use these indicator properties to assess several recent AI systems, and we discuss how future systems might implement them. Our analysis suggests that no current AI systems are conscious, but also suggests that there are no obvious technical barriers to building AI systems which satisfy these indicators.1 © 2023, CC BY-NC-SA.

关键词： Neurophysiology

infotheory: A C++/Python package for multivariate information theoretic analysis

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Candadai, Madhavun Izquierdo, Eduardo J. Program in Cognitive Science Indiana University Bloomington United States School of Informatics Computing and Engineering Indiana University Bloomington United States

This paper introduces infotheory: a package written in C++ and usable from Python and C++, for multivariate information theoretic analyses of discrete and continuous data. This package allows the user to study the relationship between components of a complex system simply from the data recorded during its operation, using the tools of information theory. It implements widely used measures such as entropy and mutual information, as well as more recent measures that arise from multivariate extensions to information theory, specifically Partial Information Decomposition. It provides an easy-to-use and flexible tool for use in research as well as pedgogical purposes to introduce students to information theory. Copyright © 2019, The Authors. All rights reserved.

关键词： Information theory

The EMPOWER blended digital intervention for relapse prevention in schizophrenia: a feasibility cluster randomised controlled trial in Scotland and Australia

学校读者我要写书评

暂无评论

The Lancet Psychiatry 2022年第6期9卷 477-486页

作者： Gumley, Andrew I Bradstreet, Simon Ainsworth, John Allan, Stephanie Alvarez-Jimenez, Mario Aucott, Lorna Birchwood, Maximillian Briggs, Andrew Bucci, Sandra Cotton, Sue M Engel, Lidia French, Paul Lederman, Reeva Lewis, Shôn Machin, Matthew MacLennan, Graeme McLeod, Hamish McMeekin, Nicola Mihalopoulos, Cathy Morton, Emma Norrie, John Schwannauer, Matthias Singh, Swaran P Sundram, Suresh Thompson, Andrew Williams, Chris Yung, Alison R Farhall, John Gleeson, John Glasgow Institute of Health and Wellbeing University of Glasgow Glasgow United Kingdom Division of Psychology and Mental Health School of Health Sciences Faculty of Biology Medicine and Health Manchester Academic Health Science Centre University of Manchester Manchester United Kingdom Division of Informatics Imaging and Data Sciences School of Health Sciences Faculty of Biology Medicine and Health Manchester Academic Health Science Centre University of Manchester Manchester United Kingdom Orygen Melbourne Melbourne VIC Australia Centre for Youth Mental Health University of Melbourne Melbourne VIC Australia School of Computing and Information Systems Faculty of Engineering and Information Technology University of Melbourne Melbourne VIC Australia Centre for Healthcare Randomised Trials (CHaRT) University of Aberdeen Aberdeen United Kingdom Centre for Mental Health and Wellbeing Research Warwick Medical School University of Warwick Warwick United Kingdom Department of Health Services Research and Policy London School of Hygiene & Tropical Medicine London United Kingdom Greater Manchester Mental Health NHS Foundation Trust Manchester United Kingdom School of Public Health and Preventive Medicine Monash University Melbourne VIC Australia Department of Psychiatry School of Clinical Sciences Monash University Melbourne VIC Australia School of Medicine Deakin University Melbourne VIC Australia Department of Psychiatry Manchester Metropolitan University Manchester United Kingdom Department of Psychiatry University of British Columbia Vancouver BC Canada The Usher Institute University of Edinburgh Edinburgh United Kingdom School for Health in Social Sciences University of Edinburgh Edinburgh United Kingdom Mental Health Program Monash Health Melbourne VIC Australia Department of Psychology and Counselling La Trobe University Melbourne VIC Australia NorthWestern Mental Health The Royal Melbourne Hospital Epping VIC Australia Healthy Brain and Mi

Background: Early warning signs monitoring by service users with schizophrenia has shown promise in preventing relapse but the quality of evidence is low. We aimed to establish the feasibility of undertaking a definitive randomised controlled trial to determine the effectiveness of a blended digital intervention for relapse prevention in schizophrenia. Methods: This multicentre, feasibility, cluster randomised controlled trial aimed to compare Early signs Monitoring to Prevent relapse in psychosis and prOmote Well-being, Engagement, and Recovery (EMPOWER) with treatment as usual in community mental health services (CMHS) in Glasgow and Melbourne. CMHS were the unit of randomisation, selected on the basis of those that probably had five or more care coordinators willing to participate. Participants were eligible if they were older than 16 years, had a schizophrenia or related diagnosis confirmed via case records, were able to provide informed consent, had contact with CMHS, and had had a relapse within the previous 2 years. Participants were randomised within stratified clusters to EMPOWER or to continue their usual approach to care. EMPOWER blended a smartphone for active monitoring of early warning signs with peer support to promote self-management and clinical triage to promote access to relapse prevention. Main outcomes were feasibility, acceptability, usability, and safety, which was assessed through face-to-face interviews. App usage was assessed via the smartphone and self-report. Primary end point was 12 months. Participants, research assistants and other team members involved in delivering the intervention were not masked to treatment conditions. Assessment of relapse was done by an independent adjudication panel masked to randomisation group. The study is registered at ISRCTN (99559262). Findings: We identified and randomised eight CMHS (six in Glasgow and two in Melbourne) comprising 47 care coordinators. We recruited 86 service users between Jan 19 and Au

关键词：

Malignant peritoneal mesothelioma interactome with 417 novel protein-protein interactions

学校读者我要写书评

暂无评论

BJC reports 2024年第1期2卷 42页

作者： Kalyani B Karunakaran Madhavi K Ganapathiraju Supercomputer Education and Research Centre Indian Institute of Science Bengaluru 560012 India. kalyanithepebble@***. Department of Biomedical Informatics School of Medicine and Intelligent Systems Program School of Computing and Information University of Pittsburgh 5607 Baum Blvd 5th Floor Pittsburgh PA 15206 USA. madhavi@pitt.edu. Carnegie Mellon University in Qatar Doha Qatar. madhavi@pitt.edu.

BACKGROUND:Malignant peritoneal mesothelioma (MPeM) is an aggressive cancer affecting the abdominal peritoneal lining and intra-abdominal organs, with a median survival of ~2.5 years. METHODS:We constructed the protein interactome of 59 MPeM-associated genes with previously known protein-protein interactions (PPIs) as well as novel PPIs predicted using our previously developed HiPPIP computational model and analysed it for transcriptomic and functional associations and for repurposable drugs. RESULTS:The MPeM interactome had over 400 computationally predicted PPIs and 4700 known PPIs. Transcriptomic evidence validated 75.6% of the genes in the interactome and 65% of the novel interactors. Some genes had tissue-specific expression in extramedullary hematopoietic sites and the expression of some genes could be correlated with unfavourable prognoses in various cancers. 39 out of 152 drugs that target the proteins in the interactome were identified as potentially repurposable for MPeM, with 29 having evidence from prior clinical trials, animal models or cell lines for effectiveness against peritoneal and pleural mesothelioma and primary peritoneal cancer. Functional modules related to chromosomal segregation, transcriptional dysregulation, IL-6 production and hematopoiesis were identified from the interactome. The MPeM interactome overlapped significantly with the malignant pleural mesothelioma interactome, revealing shared molecular pathways. CONCLUSIONS:Our findings demonstrate the utility of the interactome in uncovering biological associations and in generating clinically translatable results.

关键词：

Cascaded cross-module residual learning towards lightweight end-to-end speech coding

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Zhen, Kai Sung, Jongmo Lee, Mi Suk Beack, Seungkwon Kim, Minje Indiana University School of Informatics Computing and Engineering BloomingtonIN United States Indiana University Cognitive Science Program BloomingtonIN United States Electronics and Telecommunications Research Institute Daejeon Korea Republic of

Speech codecs learn compact representations of speech signals to facilitate data transmission. Many recent deep neural network (DNN) based end-to-end speech codecs achieve low bitrates and high perceptual quality at the cost of model complexity. We propose a cross-module residual learning (CMRL) pipeline as a module carrier with each module reconstructing the residual from its preceding modules. CMRL differs from other DNN-based speech codecs, in that rather than modeling speech compression problem in a single large neural network, it optimizes a series of less-complicated modules in a two-phase training scheme. The proposed method shows better objective performance than AMR-WB and the state-of-the-art DNNbased speech codec with a similar network architecture. As an end-to-end model, it takes raw PCM signals as an input, but is also compatible with linear predictive coding (LPC), showing better subjective quality at high bitrates than AMR-WB and OPUS. The gain is achieved by using only 0.9 million trainable parameters, a significantly less complex architecture than the other DNN-based codecs in the literature. Copyright © 2019, The Authors. All rights reserved.

关键词： Deep neural networks

On training flexible robots using deep reinforcement learning

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Dwiel, Zach Candadai, Madhavun Phielipp, Mariano Intel AI Labs BloomingtonIN United States Program in Cognitive Science School of Informatics Computing and Engineering Indiana University BloomingtonIN United States Intel AI Labs PhoenixAZ United States

The use of robotics in controlled environments has flourished over the last several decades and training robots to perform tasks using control strategies developed from dynamical models of their hardware have proven very effective. However, in many real-world settings, the uncertainties of the environment, the safety requirements and generalized capabilities that are expected of robots make rigid industrial robots unsuitable. This created great research interest into developing control strategies for flexible robot hardware for which building dynamical models are challenging. In this paper, inspired by the success of deep reinforcement learning (DRL) in other areas, we systematically study the efficacy of policy search methods using DRL in training flexible robots. Our results indicate that DRL is successfully able to learn efficient and robust policies for complex tasks at various degrees of flexibility. We also note that DRL using Deep Deterministic Policy Gradients can be sensitive to the choice of sensors and adding more informative sensors does not necessarily make the task easier to learn. Copyright © 2019, The Authors. All rights reserved.

关键词： Deep learning

A dual-staged context aggregation method towards efficient end-to-end speech enhancement

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Zhen, Kai Suk Lee, Mi Kim, Minje Indiana University Luddy School of Informatics Computing and Engineering BloomingtonIN United States Indiana University Cognitive Science Program BloomingtonIN United States Electronics and Telecommunications Research Institute Daejeon Korea Republic of

In speech enhancement, an end-to-end deep neural network converts a noisy speech signal to a clean speech directly in time domain without time-frequency transformation or mask estimation. However, aggregating contextual information from a high-resolution time domain signal with an affordable model complexity still remains challenging. In this paper, we propose a densely connected convolutional and recurrent network (DCCRN), a hybrid architecture, to enable dual-staged temporal context aggregation. With the dense connectivity and cross-component identical shortcut, DCCRN consistently outperforms competing convolutional baselines with an average STOI improvement of 0.23 and PESQ of 1.38 at three SNR levels. The proposed method is computationally efficient with only 1.38 million parameters. The generalizability performance on the unseen noise types is still decent considering its low complexity, although it is relatively weaker comparing to Wave-U-Net with 7.25 times more parameters. Copyright © 2019, The Authors. All rights reserved.

关键词： Speech enhancement

The CBIT + TMS trial: study protocol for a two-phase randomized controlled trial testing neuromodulation to augment behavior therapy for youth with chronic tics

学校读者我要写书评

暂无评论

Trials 2023年第1期24卷 439-439页

作者： Conelea, Christine Greene, Deanna J. Alexander, Jennifer Houlihan, Kerry Hodapp, Sarah Wellen, Brianna Francis, Sunday Mueller, Bryon Hendrickson, Tim Tseng, Angela Chen, Mo Fiecas, Mark Lim, Kelvin Opitz, Alexander Jacob, Suma Department of Psychiatry and Behavioral Sciences Masonic Institute for the Developing Brain University of Minnesota 2025 E. River Parkway Minneapolis 55414 MN United States Department of Cognitive Science University of California San Diego San Diego United States Department of Psychiatry and Behavioral Sciences University of Minnesota Minneapolis United States Masonic Institute for the Developing Brain University of Minnesota Informatics Institute Minneapolis United States Non-Invasive Neuromodulation Lab Brain Conditions MnDRIVE Initiative University of Minnesota Minneapolis United States Neuroscience Program Research Department Gillette Children’s Specialty Healthcare Saint Paul United States School of Public Health Division of Biostatistics University of Minnesota Minneapolis United States Department of Biomedical Engineering University of Minnesota Minneapolis United States

Background: Comprehensive Behavioral Intervention for Tics (CBIT) is a first-line treatment for tic disorders that aims to improve controllability over tics that an individual finds distressing or impairing. However, it is only effective for approximately half of patients. Supplementary motor area (SMA)-directed neurocircuitry plays a strong role in motor inhibition, and activity in this region is thought to contribute to tic expression. Targeted modulation of SMA using transcranial magnetic stimulation (TMS) may increase CBIT efficacy by improving patients' ability to implement tic controllability behaviors. Methods: The CBIT + TMS trial is a two-phase, milestone-driven early-stage randomized controlled trial. The trial will test whether augmenting CBIT with inhibitory, non-invasive stimulation of SMA with TMS modifies activity in SMA-mediated circuits and enhances tic controllability in youth ages 12–21 years with chronic tics. Phase 1 will directly compare two rTMS augmentation strategies (1 Hz rTMS vs. cTBS) vs. sham in N = 60 participants. Quantifiable, a priori “Go/No Go Criteria” guide the decision to proceed to phase 2 and the selection of the optimal TMS regimen. Phase 2 will compare the optimal regimen vs. sham and test the link between neural target engagement and clinical outcomes in a new sample of N = 60 participants. Discussion: This clinical trial is one of few to date testing TMS augmentation of therapy in a pediatric sample. The results will provide insight into whether TMS is a potentially viable strategy for enhancing CBIT efficacy and reveal potential neural and behavioral mechanisms of change. Trial registration: *** NCT04578912 . Registered on October 8, 2020. © 2023, The Author(s).

关键词： Behavior therapy Neuromodulation Pediatric Tic Tourette Transcranial magnetic stimulation

On Training Flexible Robots using Deep Reinforcement Learning

学校读者我要写书评

暂无评论

On Training Flexible Robots using Deep Reinforcement Learnin...

2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

作者： Zach Dwiel Madhavun Candadai Mariano Phielipp Intel AI Labs Bloomington IN U.S.A Program in Cognitive Science and School of Informatics Computing and Engineering at Indiana University Bloomington IN U.S.A Intel AI Labs Phoenix AZ U.S.A

The use of robotics in controlled environments has flourished over the last several decades and training robots to perform tasks using control strategies developed from dynamical models of their hardware have proven very effective. However, in many real-world settings, the uncertainties of the environment, the safety requirements and generalized capabilities that are expected of robots make rigid industrial robots unsuitable. This created great research interest in developing control strategies for flexible robot hardware for which building dynamical models are challenging. In this paper, inspired by the success of deep reinforcement learning (DRL), we systematically study the efficacy of policy search methods using DRL in training flexible robots. Our results indicate that DRL is successfully able to learn efficient and robust policies for complex tasks at various degrees of flexibility. We also note that DRL using Deep Deterministic Policy Gradients can be sensitive to the choice of sensors and adding more informative sensors does not necessarily make the task easier to learn.

关键词：