scholarly journals Fine-Tuning and Training of DenseNet for Histopathology Image Representation Using TCGA Diagnostic Slides

2021 ◽  
pp. 102032
Author(s):  
Abtin Riasatian ◽  
Morteza Babaie ◽  
Danial Maleki ◽  
Shivam Kalra ◽  
Mojtaba Valipour ◽  
...  
Keyword(s):  
Image Representation ◽  
Fine Tuning ◽  
And Training

scholarly journals Recognizing human activities from smartphone sensors using hierarchical continuous hidden Markov models

2017 ◽  
Vol 13 (1) ◽  
pp. 155014771668368 ◽  
Author(s):  
Charissa Ann Ronao ◽  
Sung-Bae Cho
Keyword(s):  
Hierarchical Structure ◽  
Markov Models ◽  
Hidden Markov ◽  
Classification Performance ◽  
Fine Tuning ◽  
Sensor Data ◽  
Model Framework ◽  
The Hierarchical Structure ◽  
Smartphone Sensors ◽  
And Training

Human activity recognition has been gaining more and more attention from researchers in recent years, particularly with the use of widespread and commercially available devices such as smartphones. However, most of the existing works focus on discriminative classifiers while neglecting the inherent time-series and continuous characteristics of sensor data. To address this, we propose a two-stage continuous hidden Markov model framework, which also takes advantage of the innate hierarchical structure of basic activities. This kind of system architecture not only enables the use of different feature subsets on different subclasses, which effectively reduces feature computation overhead, but also allows for varying number of states and iterations. Experiments show that the hierarchical structure dramatically increases classification performance. We analyze the behavior of the accelerometer and gyroscope signals for each activity through graphs, and with added fine tuning of states and training iterations, the proposed method is able to achieve an overall accuracy of up to 93.18%, which is the best performance among the state-of-the-art classifiers for the problem at hand.

scholarly journals Beyond Network Pruning: a Joint Search-and-Training Approach

2020 ◽  
Author(s):  
Xiaotong Lu ◽  
Han Huang ◽  
Weisheng Dong ◽  
Xin Li ◽  
Guangming Shi
Keyword(s):  
Random Perturbation ◽  
Search Space ◽  
Fine Tuning ◽  
Superior Performance ◽  
Network Pruning ◽  
Training Approach ◽  
Target Network ◽  
Tuning Strategy ◽  
Coarse To Fine ◽  
And Training

Network pruning has been proposed as a remedy for alleviating the over-parameterization problem of deep neural networks. However, its value has been recently challenged especially from the perspective of neural architecture search (NAS). We challenge the conventional wisdom of pruning-after-training by proposing a joint search-and-training approach that directly learns a compact network from the scratch. By treating pruning as a search strategy, we present two new insights in this paper: 1) it is possible to expand the search space of networking pruning by associating each filter with a learnable weight; 2) joint search-and-training can be conducted iteratively to maximize the learning efficiency. More specifically, we propose a coarse-to-fine tuning strategy to iteratively sample and update compact sub-network to approximate the target network. The weights associated with network filters will be accordingly updated by joint search-and-training to reflect learned knowledge in NAS space. Moreover, we introduce strategies of random perturbation (inspired by Monte Carlo) and flexible thresholding (inspired by Reinforcement Learning) to adjust the weight and size of each layer. Extensive experiments on ResNet and VGGNet demonstrate the superior performance of our proposed method on popular datasets including CIFAR10, CIFAR100 and ImageNet.

scholarly journals Optimising recruitment to the HAND-1 RCT feasibility study: integration of the QuinteT Recruitment Intervention (QRI)

2020 ◽  
Vol 6 (1) ◽  
Author(s):  
Samantha Husbands ◽  
Daisy Elliott ◽  
Tim R. C. Davis ◽  
Jane M. Blazeby ◽  
Eleanor F. Harrison ◽  
...  
Keyword(s):  
Feasibility Study ◽  
Good Practice ◽  
Feasibility Studies ◽  
Fine Tuning ◽  
Screening Methods ◽  
Main Stage ◽  
Main Trial ◽  
Recruitment Challenges ◽  
Set Up ◽  
And Training

Abstract Background Recruitment to randomised controlled trials (RCTs) can be challenging, with most trials not reaching recruitment targets. Randomised feasibility studies can be set up prior to a main trial to identify and overcome recruitment obstacles. This paper reports on an intervention—the QuinteT Recruitment Intervention (QRI)—to optimise recruitment within a randomised feasibility study of surgical treatments for patients with Dupuytren’s contracture (the HAND-1 study). Methods The QRI was introduced in 2-phases: phase 1 sought to understand the recruitment challenges by interviewing trial staff, scrutinising screening logs and analysing audio-recorded patient consultations; in phase 2 a tailored plan of action consisting of recruiter feedback and training was delivered to address the identified challenges. Results Two key recruitment obstacles emerged: (1) issues with the recruitment pathway, in particular methods to identify potentially eligible patients and (2) equipoise of recruiters and patients. These were addressed by liaising with centres to share good practice and refine their pathway and by providing bespoke feedback and training on consent discussions to individual recruiters and centres whilst recruitment was ongoing. The HAND-1 study subsequently achieved its recruitment target. Conclusions Transferable lessons learnt from the QRI in the feasibility study will be implemented in the definitive RCT, enabling a “head start” in the tackling of wider issues around screening methods and consent discussions in the set up/early recruitment study phases, with ongoing QRI addressing specific issues with new centres and recruiters. Findings from this study are likely to be relevant to other surgical and similar trials that are anticipated to encounter issues around patient and recruiter equipoise of treatments and variation in recruitment pathways across centres. The study also highlights the value of feasibility studies in fine-tuning design and conduct issues for definitive RCTs. Embedding a QRI in an RCT, at feasibility or main stage, offers an opportunity for a detailed and nuanced understanding of key recruitment challenges and the chance to address them in “real-time” as recruitment proceeds.

scholarly journals Threshold-Based Hierarchical Clustering for Person Re-Identification

Entropy ◽  
2021 ◽  
Vol 23 (5) ◽  
pp. 522
Author(s):  
Minhui Hu ◽  
Kaiwei Zeng ◽  
Yaohua Wang ◽  
Yang Guo
Keyword(s):  
Hierarchical Clustering ◽  
Domain Adaptation ◽  
State Of The Art ◽  
Fine Tuning ◽  
Single Sample ◽  
Clustering Methods ◽  
Clustering Method ◽  
Cluster Number ◽  
Unsupervised Domain Adaptation ◽  
And Training

Unsupervised domain adaptation is a challenging task in person re-identification (re-ID). Recently, cluster-based methods achieve good performance; clustering and training are two important phases in these methods. For clustering, one major issue of existing methods is that they do not fully exploit the information in outliers by either discarding outliers in clusters or simply merging outliers. For training, existing methods only use source features for pretraining and target features for fine-tuning and do not make full use of all valuable information in source datasets and target datasets. To solve these problems, we propose a Threshold-based Hierarchical clustering method with Contrastive loss (THC). There are two features of THC: (1) it regards outliers as single-sample clusters to participate in training. It well preserves the information in outliers without setting cluster number and combines advantages of existing clustering methods; (2) it uses contrastive loss to make full use of all valuable information, including source-class centroids, target-cluster centroids and single-sample clusters, thus achieving better performance. We conduct extensive experiments on Market-1501, DukeMTMC-reID and MSMT17. Results show our method achieves state of the art.

scholarly journals TRP: Trained Rank Pruning for Efficient Deep Neural Networks

2020 ◽  
Author(s):  
Yuhui Xu ◽  
Yuxi Li ◽  
Shuai Zhang ◽  
Wei Wen ◽  
Botao Wang ◽  
...  
Keyword(s):  
Fine Tuning ◽  
Low Rank ◽  
Low Rank Approximation ◽  
Performance Loss ◽  
Approximation Errors ◽  
Rank Approximation ◽  
Rank Decomposition ◽  
Rank Constraints ◽  
Low Rank Decomposition ◽  
And Training

To enable DNNs on edge devices like mobile phones, low-rank approximation has been widely adopted because of its solid theoretical rationale and efficient implementations. Several previous works attempted to directly approximate a pre-trained model by low-rank decomposition; however, small approximation errors in parameters can ripple over a large prediction loss. As a result, performance usually drops significantly and a sophisticated effort on fine-tuning is required to recover accuracy. Apparently, it is not optimal to separate low-rank approximation from training. Unlike previous works, this paper integrates low rank approximation and regularization into the training process. We propose Trained Rank Pruning (TRP), which alternates between low rank approximation and training. TRP maintains the capacity of the original network while imposing low-rank constraints during training. A nuclear regularization optimized by stochastic sub-gradient descent is utilized to further promote low rank in TRP. The TRP trained network inherently has a low-rank structure, and is approximated with negligible performance loss, thus eliminating the fine-tuning process after low rank decomposition. The proposed method is comprehensively evaluated on CIFAR-10 and ImageNet, outperforming previous compression methods using low rank approximation.

Managing Drought in Australia's Rangelands: Four Weddings and a Funeral.

1996 ◽  
Vol 18 (2) ◽  
pp. 292 ◽  
Author(s):  
R Buxton ◽  
MS Smith
Keyword(s):  
Cash Flow ◽  
Climatic Variability ◽  
Fine Tuning ◽  
Property Management ◽  
Management Skills ◽  
Short Term ◽  
Management Options ◽  
Systematic Analysis ◽  
And Training

We report some of the findings of a project called 'DroughtPlan', which has involved close collaboration with pastoralists throughout the Australian rangelands. There were three general areas related to property management where producers felt better information and training could help them cope with climatic variability. These were strategic management of long-term stocking levels, tactical management of stock numbers between years, and business management skills. A comprehensive series of representative studies linking herd biology with economic outcomes was undertaken on these topics with pastoralists in different regions of the rangelands. Some of the studies considered most important by pastoralists are reported here. These demonstrate that: (i) a reduction in current stocking levels can often improve cash flow; (ii) small adjustments in livestock selling tactics during drought can have large financial ramifications; (iii) it is financially advantageous to build stock numbers up quickly after a drought, even though this may conflict with longer-term environmental values; (iv) while diversification can provide financial rewards, these could be matched by small improvements in the biological rates of the livestock; and, (v) fine-tuning of the existing pastoral enterprise can provide a less risky means of increasing cash flow and reducing its variability than does diversification. Four of the examples indicate that better use of information can help both profitability and sustainability; the fifth suggests that the interests of short-term profitability are in conflict with long-term land conservation goals, as assessed by many pastoralists. These studies highlight the value of linking producer knowledge with a systematic analysis framework, as well as the vital importance of incorporating the effects of climatic variability, when assessing the value of different management options.

scholarly journals Recognition of 3D Shapes Based on 3V-DepthPano CNN

2020 ◽  
Vol 2020 ◽  
pp. 1-11 ◽  
Author(s):  
Junjie Yin ◽  
Ningning Huang ◽  
Jing Tang ◽  
Meie Fang
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Shape Descriptor ◽  
Recognition System ◽  
Fine Tuning ◽  
Training Time ◽  
3D Shape ◽  
3D Shapes ◽  
And Training

This paper proposes a convolutional neural network (CNN) with three branches based on the three-view drawing principle and depth panorama for 3D shape recognition. The three-view drawing principle provides three key views of a 3D shape. A depth panorama contains the complete 2.5D information of each view. 3V-DepthPano CNN is a CNN system with three branches designed for depth panoramas generated from the three key views. This recognition system, i.e., 3V-DepthPano CNN, applies a three-branch convolutional neural network to aggregate the 3D shape depth panorama information into a more compact 3D shape descriptor to implement the classification of 3D shapes. Furthermore, we adopt a fine-tuning technique on 3V-DepthPano CNN and extract shape features to facilitate the retrieval of 3D shapes. The proposed method implements a good tradeoff state between higher accuracy and training time. Experiments show that the proposed 3V-DepthPano CNN with 3 views obtains approximate accuracy to MVCNN with 12/80 views. But the 3V-DepthPano CNN frame takes much shorter time to obtain depth panoramas and train the network than MVCNN. It is superior to all other existing advanced methods for both classification and shape retrieval.

An Exploratory Study on Augmentative and Alternative Communication Classroom Integration Training for Special Educators Who Work With Children With Complex Communication Needs in the Bahamas

2019 ◽  
Vol 4 (5) ◽  
pp. 991-1016
Author(s):  
Shameka Stanford ◽  
Ovetta Harris
Keyword(s):  
Augmentative And Alternative Communication ◽  
Special Educators ◽  
Communication Disorders ◽  
Alternative Communication ◽  
The Bahamas ◽  
Complex Communication Needs ◽  
English Speaking ◽  
Communication Needs ◽  
Classroom Integration ◽  
And Training

Purpose In 2011, the United Nations estimated there were between 180 and 220 million youth with disabilities living around the world, and 80% of them resided in developing countries. Over the last 6 years, this number has increased significantly, and now, over 1 million people live in the Caribbean with some form of disability such as communication disorders resulting in complex communication needs (CCN). Method This publication discusses the benefits of an exploratory, descriptive, nonexperimental study on augmentative and alternative communication (AAC) classroom integration training for 8 special educators in the Bahamas who work with children with CCN. Results The results of this study revealed that 100% of the participants reported the study to be effective in increasing their knowledge and skill in the area of implementing AAC into their classrooms, enhancing their ability to team teach and incorporate AAC opportunities for all students with CCN within their classrooms, and increasing their knowledge and skill overall in the areas of AAC and CCN. Conclusion The findings highlight an important area of potential professional development and training that can be replicated in other English-speaking Caribbean territories focused on AAC classroom integration training program for special educators who teach students with CCN.

Sign in / Sign up

Export Citation Format

Share Document