scholarly journals Targeted transfer learning to improve performance in small medical physics datasets

2020 ◽  
Author(s):  
Miguel Romero ◽  
Yannet Interian ◽  
Timothy Solberg ◽  
Gilmer Valdes
Keyword(s):  
Transfer Learning ◽  
Medical Physics ◽  
Improve Performance

scholarly journals The Need for Transfer Learning in CRISPR-Cas Off-Target Scoring

2021 ◽  
Author(s):  
Pavan K Kota ◽  
Yidan Pan ◽  
Hoang-Anh Vu ◽  
Mingming Cao ◽  
Richard G Baraniuk ◽  
...  
Keyword(s):  
Transfer Learning ◽  
Benchmark Dataset ◽  
Model Complexity ◽  
Improve Performance ◽  
Rna Sequences ◽  
Guide Rna ◽  
Learning Techniques ◽  
Scalable Design ◽  
Biological Differences

The scalable design of safe guide RNA sequences for CRISPR gene editing depends on the computational "scoring" of DNA locations that may be edited. As there is no widely accepted benchmark dataset to compare scoring models, we present a curated "TrueOT" dataset that contains thoroughly validated datapoints to best reflect the properties of in vivo editing. Many existing models are trained on data from high throughput assays. We hypothesize that such models may suboptimally transfer to the low throughput data in TrueOT due to fundamental biological differences between proxy assays and in vivo behavior. We developed new Siamese convolutional neural networks, trained them on a proxy dataset, and compared their performance against existing models on TrueOT. Our simplest model with a single convolutional and pooling layer surprisingly exhibits state-ofthe-art performance on TrueOT. Adding subsequent layers improves performance on the proxy dataset while compromising performance on TrueOT. We demonstrate that model complexity can only improve performance on TrueOT if transfer learning techniques are employed. These results suggest an urgent need for the CRISPR community to agree upon a benchmark dataset such as TrueOT and highlight that various sources of CRISPR data cannot be assumed to be equivalent. Our codebase and datasets are available on GitHub at github.com/baolab-rice/CRISPR_OT_scoring.

scholarly journals The Case for Case-Based Transfer Learning

AI Magazine ◽  
2011 ◽  
Vol 32 (1) ◽  
pp. 54 ◽  
Author(s):  
Matthew Klenk ◽  
David W. Aha ◽  
Matt Molineaux
Keyword(s):  
Problem Solving ◽  
Transfer Learning ◽  
Future Research ◽  
Case Based Reasoning ◽  
Learning Method ◽  
Improve Performance ◽  
Learning Methods ◽  
Problem Solving Process ◽  
Case Based ◽  
Solution Transfer

Case-based reasoning (CBR) is a problem-solving process in which a new problem is solved by retrieving a similar situation and reusing its solution. Transfer learning occurs when, after gaining experience from learning how to solve source problems, the same learner exploits this experience to improve performance and/or learning on target problems. In transfer learning, the differences between the source and target problems characterize the transfer distance. CBR can support transfer learning methods in multiple ways. We illustrate how CBR and transfer learning interact and characterize three approaches for using CBR in transfer learning: (1) as a transfer learning method, (2) for problem learning, and (3) to transfer knowledge between sets of problems. We describe examples of these approaches from our own and related work and discuss applicable transfer distances for each. We close with conclusions and directions for future research applying CBR to transfer learning.

scholarly journals Bangladeshi Native Vehicle Classification Based on Transfer Learning with Deep Convolutional Neural Network

Sensors ◽  
2021 ◽  
Vol 21 (22) ◽  
pp. 7545
Author(s):  
Md Mahibul Hasan ◽  
Zhijie Wang ◽  
Muhammad Ather Iqbal Hussain ◽  
Kaniz Fatima
Keyword(s):  
Transfer Learning ◽  
Data Augmentation ◽  
Intelligent Transportation System ◽  
Vehicle Classification ◽  
Improve Performance ◽  
Vehicle Type ◽  
Proposed Model ◽  
Vehicle Type Classification ◽  
Type Classification

Vehicle type classification plays an essential role in developing an intelligent transportation system (ITS). Based on the modern accomplishments of deep learning (DL) on image classification, we proposed a model based on transfer learning, incorporating data augmentation, for the recognition and classification of Bangladeshi native vehicle types. An extensive dataset of Bangladeshi native vehicles, encompassing 10,440 images, was developed. Here, the images are categorized into 13 common vehicle classes in Bangladesh. The method utilized was a residual network (ResNet-50)-based model, with extra classification blocks added to improve performance. Here, vehicle type features were automatically extracted and categorized. While conducting the analysis, a variety of metrics was used for the evaluation, including accuracy, precision, recall, and F1 − Score. In spite of the changing physical properties of the vehicles, the proposed model achieved progressive accuracy. Our proposed method surpasses the existing baseline method as well as two pre-trained DL approaches, AlexNet and VGG-16. Based on result comparisons, we have seen that, in the classification of Bangladeshi native vehicle types, our suggested ResNet-50 pre-trained model achieves an accuracy of 98.00%.

scholarly journals Deep convolutional neural network based medical image classification for disease diagnosis

2019 ◽  
Vol 6 (1) ◽  
Author(s):  
Samir S. Yadav ◽  
Shivajirao M. Jadhav
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Image Classification ◽  
Transfer Learning ◽  
Medical Image ◽  
Data Augmentation ◽  
Support Vector ◽  
Improve Performance ◽  
Classification Tasks ◽  
Medical Image Classification

AbstractMedical image classification plays an essential role in clinical treatment and teaching tasks. However, the traditional method has reached its ceiling on performance. Moreover, by using them, much time and effort need to be spent on extracting and selecting classification features. The deep neural network is an emerging machine learning method that has proven its potential for different classification tasks. Notably, the convolutional neural network dominates with the best results on varying image classification tasks. However, medical image datasets are hard to collect because it needs a lot of professional expertise to label them. Therefore, this paper researches how to apply the convolutional neural network (CNN) based algorithm on a chest X-ray dataset to classify pneumonia. Three techniques are evaluated through experiments. These are linear support vector machine classifier with local rotation and orientation free features, transfer learning on two convolutional neural network models: Visual Geometry Group i.e., VGG16 and InceptionV3, and a capsule network training from scratch. Data augmentation is a data preprocessing method applied to all three methods. The results of the experiments show that data augmentation generally is an effective way for all three algorithms to improve performance. Also, Transfer learning is a more useful classification method on a small dataset compared to a support vector machine with oriented fast and rotated binary (ORB) robust independent elementary features and capsule network. In transfer learning, retraining specific features on a new target dataset is essential to improve performance. And, the second important factor is a proper network complexity that matches the scale of the dataset.

scholarly journals Autonomous Task Sequencing for Customized Curriculum Design in Reinforcement Learning

2017 ◽  
Author(s):  
Sanmit Narvekar ◽  
Jivko Sinapov ◽  
Peter Stone
Keyword(s):  
Reinforcement Learning ◽  
Transfer Learning ◽  
Markov Decision Process ◽  
Curriculum Design ◽  
Optimal Policy ◽  
Decision Process ◽  
Improve Performance ◽  
Task Sequencing ◽  
Action Capabilities ◽  
Markov Decision

Transfer learning is a method where an agent reuses knowledge learned in a source task to improve learning on a target task. Recent work has shown that transfer learning can be extended to the idea of curriculum learning, where the agent incrementally accumulates knowledge over a sequence of tasks (i.e. a curriculum). In most existing work, such curricula have been constructed manually. Furthermore, they are fixed ahead of time, and do not adapt to the progress or abilities of the agent. In this paper, we formulate the design of a curriculum as a Markov Decision Process, which directly models the accumulation of knowledge as an agent interacts with tasks, and propose a method that approximates an execution of an optimal policy in this MDP to produce an agent-specific curriculum. We use our approach to automatically sequence tasks for 3 agents with varying sensing and action capabilities in an experimental domain, and show that our method produces curricula customized for each agent that improve performance relative to learning from scratch or using a different agent's curriculum.

scholarly journals Omnidirectional Transfer for Quasilinear Lifelong Learning

2021 ◽  
Author(s):  
Jayanta Dey ◽  
Joshua Vogelstein ◽  
Hayden Helm ◽  
Will Levine ◽  
Ronak Mehta ◽  
...  
Keyword(s):  
Lifelong Learning ◽  
Transfer Learning ◽  
Learning Algorithms ◽  
Image Data ◽  
Real Data ◽  
Improve Performance ◽  
Special Cases ◽  
Current Task ◽  
Biological Learning ◽  
Using Data

Abstract In biological learning, data are used to improve performance not only on the current task, but also on previously encountered and as yet unencountered tasks. In contrast, classical machine learning starts from a blank slate, or tabula rasa, using data only for the single task at hand. While typical transfer learning algorithms can improve performance on future tasks, their performance on prior tasks degrades upon learning new tasks (called catastrophic forgetting). Many recent approaches for continual or lifelong learning have attempted to maintain performance given new tasks. But striving to avoid forgetting sets the goal unnecessarily low: the goal of lifelong learning, whether biological or artificial, should be to improve performance on all tasks (including past and future) with any new data. We propose omnidirectional transfer learning algorithms, which includes two special cases of interest: decision forests and deep networks. Our key insight is the development of the omni-voter layer, which ensembles representations learned independently on all tasks to jointly decide how to proceed on any given new data point, thereby improving performance on both past and future tasks. Our algorithms demonstrate omnidirectional transfer in a variety of simulated and real data scenarios, including tabular data, image data, spoken data, and adversarial tasks. Moreover, they do so with quasilinear space and time complexity.

scholarly journals Mixing and Matching Emotion Frameworks: Investigating Cross-Framework Transfer Learning for Dutch Emotion Detection

Electronics ◽  
2021 ◽  
Vol 10 (21) ◽  
pp. 2643
Author(s):  
Luna De Bruyne ◽  
Orphée De Clercq ◽  
Véronique Hoste
Keyword(s):  
Regression Model ◽  
Transfer Learning ◽  
Classification Task ◽  
Emotion Detection ◽  
Emotion Classification ◽  
Improve Performance ◽  
Field Of Study ◽  
Broad Application ◽  
Meta Learning ◽  
Application Potential

Emotion detection has become a growing field of study, especially seeing its broad application potential. Research usually focuses on emotion classification, but performance tends to be rather low, especially when dealing with more advanced emotion categories that are tailored to specific tasks and domains. Therefore, we propose the use of the dimensional emotion representations valence, arousal and dominance (VAD), in an emotion regression task. Firstly, we hypothesize that they can improve performance of the classification task, and secondly, they might be used as a pivot mechanism to map towards any given emotion framework, which allows tailoring emotion frameworks to specific applications. In this paper, we examine three cross-framework transfer methodologies: multi-task learning, in which VAD regression and classification are learned simultaneously; meta-learning, where VAD regression and emotion classification are learned separately and predictions are jointly used as input for a meta-learner; and a pivot mechanism, which converts the predictions of the VAD model to emotion classes. We show that dimensional representations can indeed boost performance for emotion classification, especially in the meta-learning setting (up to 7% macro F1-score compared to regular emotion classification). The pivot method was not able to compete with the base model, but further inspection suggests that it could be efficient, provided that the VAD regression model is further improved.

scholarly journals Attention U-Net ensemble for interpretable polyp and instrument segmentation

2021 ◽  
Vol 1 (1) ◽  
pp. 47-49
Author(s):  
Michael Yeung
Keyword(s):  
Deep Learning ◽  
Transfer Learning ◽  
Colorectal Polyps ◽  
Test Time ◽  
Connected Component ◽  
Improve Performance ◽  
Connected Component Labeling ◽  
Gastrointestinal Pathology ◽  
Computer Aided ◽  
Salient Regions

The difficulty associated with screening and treating colorectal polyps alongside other gastrointestinal pathology presents an opportunity to incorporate computer-aided systems. This paper develops a deep learning pipeline that accurately segments colorectal polyps and various instruments used during endoscopic procedures. To improve transparency, we leverage the Attention U-Net architecture, enabling visualisation of the attention coefficients to identify salient regions. Moreover, we improve performance by incorporating transfer learning using a pre-trained encoder, together with test-time augmentation, softmax averaging, softmax thresholding and connected component labeling to further refine predictions.

Replication of the Superstition and Performance Study by

2014 ◽  
Vol 45 (3) ◽  
pp. 239-245 ◽  
Author(s):  
Robert J. Calin-Jageman ◽  
Tracy L. Caldwell
Keyword(s):  
Task Difficulty ◽  
Statistical Power ◽  
Meta Analysis ◽  
A Priori ◽  
Significant Heterogeneity ◽  
Performance Study ◽  
Improve Performance ◽  
Research Designs ◽  
Series Of Experiments ◽  
And Performance

A recent series of experiments suggests that fostering superstitions can substantially improve performance on a variety of motor and cognitive tasks ( Damisch, Stoberock, & Mussweiler, 2010 ). We conducted two high-powered and precise replications of one of these experiments, examining if telling participants they had a lucky golf ball could improve their performance on a 10-shot golf task relative to controls. We found that the effect of superstition on performance is elusive: Participants told they had a lucky ball performed almost identically to controls. Our failure to replicate the target study was not due to lack of impact, lack of statistical power, differences in task difficulty, nor differences in participant belief in luck. A meta-analysis indicates significant heterogeneity in the effect of superstition on performance. This could be due to an unknown moderator, but no effect was observed among the studies with the strongest research designs (e.g., high power, a priori sampling plan).

Sign in / Sign up

Export Citation Format

Share Document