Filters








29,642 Hits in 5.7 sec

Learning To Recognize Procedural Activities with Distant Supervision [article]

Xudong Lin, Fabio Petroni, Gedas Bertasius, Marcus Rohrbach, Shih-Fu Chang, Lorenzo Torresani
2022 arXiv   pre-print
While step annotations could enable the training of models to recognize the individual steps of procedural activities, existing large-scale datasets in this area do not include such segment labels due  ...  To address this issue, we propose to automatically identify steps in instructional videos by leveraging the distant supervision of a textual knowledge base (wikiHow) that includes detailed descriptions  ...  Acknowledgments Thanks to Karl Ridgeway, Michael Iuzzolino, Jue Wang, Noureldien Hussein, and Effrosyni Mavroudi for valuable discussions.  ... 
arXiv:2201.10990v3 fatcat:ghjybqtitjf5thlsgoknmpgf2a

Robust Distant Supervision Relation Extraction via Deep Reinforcement Learning [article]

Pengda Qin, Weiran Xu, William Yang Wang
2018 arXiv   pre-print
To do this, our paper describes a radical solution---We explore a deep reinforcement learning strategy to generate the false-positive indicator, where we automatically recognize false positives for each  ...  Distant supervision has become the standard method for relation extraction.  ...  In order to overcome this problem, we adopt a supervised learning procedure to pre-train our policy network, in which way to provide a general learning direction for our policy-based agent.  ... 
arXiv:1805.09927v1 fatcat:flxi3fggendm3gbaxmyeowczwy

Robust Distant Supervision Relation Extraction via Deep Reinforcement Learning

Pengda Qin, Weiran Xu, William Yang Wang
2018 Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)  
To do this, our paper describes a radical solution-We explore a deep reinforcement learning strategy to generate the false-positive indicator, where we automatically recognize false positives for each  ...  Distant supervision has become the standard method for relation extraction.  ...  Reinforcement Learning for Distant Supervision We introduce a performance-driven, policy-based reinforcement learning method to heuristically recognize false positive samples.  ... 
doi:10.18653/v1/p18-1199 dblp:conf/acl/WangXQ18a fatcat:tnk6kizs6vel5ntla42ypi36wq

Call for Papers! Special Issue on Models of Discovery: Mining the Scientific Literature

Neil R. Smalheiser, Aaron M. Cohen
2017 Journal of Data and Information Science  
analysis, argumentation analysis, or recognizing humor and sarcasm; • Use of outside information to guide discovery, such as bibliometrics, distant supervision, or cross-corpus information; • Use of implicit  ...  All manuscripts submitted to the special issue will be reviewed in accordance with standard journal procedures and be finally evaluated by the Issue Editors and Co-EICs.  ... 
doi:10.1515/jdis-2017-0010 fatcat:kxgl73xgk5dydfshbj3eqm4zem

Denoising Distantly Supervised Named Entity Recognition via a Hypergeometric Probabilistic Model [article]

Wenkai Zhang, Hongyu Lin, Xianpei Han, Le Sun, Huidan Liu, Zhicheng Wei, Nicholas Jing Yuan
2021 arXiv   pre-print
Denoising is the essential step for distant supervision based named entity recognition.  ...  In this paper, we propose Hypergeometric Learning (HGL), a denoising algorithm for distantly supervised NER that takes both noise distribution and instance-level confidence into consideration.  ...  the mislabeling in distant supervision by regarding it as a positive-unlabeled learning problem and reshapes the training procedure with a regularizer considers the mislabeling ratio.  ... 
arXiv:2106.09234v1 fatcat:jjnpadhybnd4xbs5huvwwbr7yy

Survey on Human Activity Recognition using Smartphone

Adeeba KH, Laheeb Ibrahim
2021 ˜Al-œRafidain journal for computer sciences and mathematics  
activities can be recognized using interactive sensors that are affected by human movement.  ...  The field of Human Activity Recognition (HAR) is an active research field in which methods are being developed to understand human behavior by interpreting features obtained from various sources, these  ...  activity by which systems are grouped with respect to the activities they are able to identify (e.g. locomotion and ADLs); the learning procedure which is related to the type of algorithms used for learning  ... 
doi:10.33899/csmj.2021.168253 fatcat:odwqb4rar5e3bhtuxeaq4hfzpm

SAIL (UNIST) team Participation in KBP 2016 Cold-Start Slot filling

Seongwoo Lim, Jiyeon Han, Kyowoon Lee, Jaesik Choi
2016 Text Analysis Conference  
We will use Distant Supervision (DS), widely used to this problem, to extract feature from unstructured text and map sentence to feature space.  ...  In this case, we use bidirectional Long Short-Term Memory (LSTM) to find the answer.  ...  First, relation extraction using Distant Supervision, second, Multi-Instance Multi-Label Learning, third, Recurrent Neural Network.  ... 
dblp:conf/tac/LimHLC16 fatcat:obs5h3ufzzdvpfwnv23bqbguki

ESC Core Curriculum for the Cardiologist

Felix C Tanner, Nicolas Brooks, Kevin F Fox, Lino Gonçalves, Peter Kearney, Lampros Michalis, Agnès Pasquet, Susanna Price, Eric Bonnefoy, Mark Westwood, Chris Plummer, Paulus Kirchhof (+1 others)
2020 European Heart Journal  
Perform a cardiological consultation Each EPA is assigned an expected level of independence from level 1 to level 5: (1) Observe; (2) Direct supervision; (3) Indirect supervision; (4) Distant supervision  ...  ; and (5) Able to teach (no supervision).  ...  Distant supervision 2.5.  ... 
doi:10.1093/eurheartj/ehaa641 pmid:32862226 fatcat:76juxv3h4rcfpkvssiddekxsmu

2020 European Society of Cardiology Core Curriculum for the Cardiologist. Is it time for change in the Portuguese cardiology training program?

Kevin Domingues, Lino Gonçalves, Maria João Vidigal, Felix C. Tanner, Miguel Mendes
2021 Revista Portuguesa de Cardiologia  
performed with distant supervision (level 4), and being able to perform the activity independently and supervise others (level 5).  ...  of independence for each entrustable professional activities and each investigation, from level 1 to level 5: (1) Observe; (2) Direct supervision; (3) Indirect supervision; (4) Distant supervision; and  ... 
doi:10.1016/j.repc.2021.03.008 pmid:34332799 fatcat:iclxcxffxzfgbc3je37ictwkk4

Tinker, Tailor, Supervisor, Spy: Lessons Learned from Distant Supervision

Elizabeth Currin, Stephanie Schroeder, Elizabeth Bondy, Brittany Castanheira
2019 Journal of Educational Supervision  
out the distant supervision model with fidelity.  ...  to support intern learning.  ...  Eager to learn and to excel in her doctoral studies, and fresh out of an elementary classroom, Anna stepped into distant supervision with purpose and enthusiasm.  ... 
doi:10.31045/jes.2.1.5 fatcat:gbz4qxi3x5akzaagsf7pcyrcmy

Syncing Our Cycles: An Inquiry-Based Coaching Model for Distant Supervision

Stephanie Schroeder, University of Florida, Elizabeth Currin, University of Florida
2019 Journal of Practitioner Research  
In response to calls for a reconceptualized approach to pre-service teacher supervision, we propose a model of distant supervision for teacher candidates that blends two evidence-based professional development  ...  Citing the dilemmas inherent in distant supervision, we argue that this hybrid coaching/inquiry model of student teaching supervision is more suitable to supervision at a distance than coaching or inquiry  ...  Indeed, we were not just learning to supervise at a distance, we were learning to coach at a distance.  ... 
doi:10.5038/2379-9951.4.1.1098 fatcat:f7miawj4dzc6zjakx63ojlatfu

University-School Cluster as an Innovative Form for Teachers' Professional Competence Development
Университетско-школьный кластер — среда развития профессиональных компетенций учителя

Galina Volodina, Alla Obolonskaya, Tatyana Ratt
2014 Educational Studies Moscow  
The paper discloses relevancy, aim, scope and technologies of experimental activity conducted by the HSE-Perm to set up the University-School Cluster as an innovative form of enhancing Perm Krai teachers  ...  Successful implementation of programs was largely provided by the distant working format applied in the cluster, with the focus on individual learning mode for each teacher supervised by a tutor.  ...  resources competitions. • Social Leader School's activities designed to develop the key supra-subject learning competencies (associated with social cul- to solve acute problems of modern-day teaching  ... 
doi:10.17323/1814-9545-2014-1-46-63 fatcat:ctadmfl23nayddidjbuk7srbr4

Mining the literature for genes associated with placenta-mediated maternal diseases

Laritza M Rodriguez, Stephanie M Morrison, Kathleen Greenberg, Dina Demner Fushman
2018 AMIA Annual Symposium Proceedings  
We developed three baseline approaches to finding sentences containing this information: one based on supervised machine learning (ML) and two based on distant supervision: 1) using automated detection  ...  We demonstrate that distant supervision approaches could be used at the expense of missing up to 15% of relevant documents.  ...  We explored two distant supervision approaches on the same collection as an alternative to manual annotation.  ... 
pmid:29854219 pmcid:PMC5977710 fatcat:j5ry7inw5zejtpamvt3qzceqgu

Integration of Entrustable Professional Activities with the Milestones for Emergency Medicine Residents

Danielle Hart, Douglas Franzen, Michael Beeson, Rahul Bhat, Miriam Kulkarni, Lorraine Thibodeau, Moshe Weizberg, Susan Promes
2018 Western Journal of Emergency Medicine  
We recognize that individual programs may have additional focus areas or work-based activities they want their trainees to achieve by graduation; therefore, programs are also encouraged to create additional  ...  While competencies describe characteristics of the person, the entrustable professional activities (EPAs) concept refers to work-related activities.  ...  ACKNOWLEDGEMENTS For assistance with the early phases of this project: Patrick Brunett, MD, Barry Brenner, MD, Lori Winston, MD.  ... 
doi:10.5811/westjem.2018.11.38912 pmid:30643599 pmcid:PMC6324698 fatcat:vefjam52ize43gtqct5dktsbyu

Tracking-based semi-supervised learning

Alex Teichman, Sebastian Thrun
2012 The international journal of robotics research  
In this paper, we consider a semi-supervised approach to the problem of track classification in dense 3D range data.  ...  When given only three hand-labeled training tracks of each object class, semi-supervised performance is comparable to that of the fully-supervised equivalent which uses thousands of hand-labeled training  ...  We refer to this method as CSSL-RR, for continuous semi-supervised learning with response relearning. This procedure is described in Algorithm 2.  ... 
doi:10.1177/0278364912442751 fatcat:7oiuczcprbbnld3danch626yei
« Previous Showing results 1 — 15 out of 29,642 results