A Survey: Limited Data Problem and Strategy of Reinforcement Learning

2021 ◽  
pp. 471-481
Author(s):  
Zhe Yang ◽  
Fengge Wu ◽  
Junsuo Zhao
Author(s):  
Vincent Francois-Lavet ◽  
Guillaume Rabusseau ◽  
Joelle Pineau ◽  
Damien Ernst ◽  
Raphael Fonteneau

When an agent has limited information on its environment, the suboptimality of an RL algorithm can be decomposed into the sum of two terms: a term related to an asymptotic bias (suboptimality with unlimited data) and a term due to overfitting (additional suboptimality due to limited data). In the context of reinforcement learning with partial observability, this paper provides an analysis of the tradeoff between these two error sources. In particular, our theoretical analysis formally characterizes how a smaller state representation increases the asymptotic bias while decreasing the risk of overfitting.


Author(s):  
Gerd Heilemann ◽  
Mark Matthewman ◽  
Peter Kuess ◽  
Gregor Goldner ◽  
Joachim Widder ◽  
...  

2021 ◽  
Vol 13 (1) ◽  
pp. 26-34
Author(s):  
Gábor Szűcs ◽  
Marcell Németh

The research topic presented in this paper belongs to small training data problem in machine learning (especially in deep learning), it intends to help the work of those working in medicine by analyzing pathological X-ray recordings, using only very few images. This scenario is a particularly hot issue nowadays: how could a new disease for which only limited data are available be diagnosed using features of previous diseases? In this problem, so-called few-shot learning, the difficulty of the classification task is to learn the unique feature characteristics associated with the classes. Although there are solutions, but if the images come from different views, they will not handle these views well. We proposed an improved method, so-called Double-View Matching Network (DVMN based on the deep neural network), which solves the few-shot learning problem as well as the different views of the pathological recordings in the images. The main contribution of this is the convolutional neural network for feature extraction and handling the multi-view in image representation. Our method was tested in the classification of images showing unknown COVID-19 symptoms in an environment designed for learning a few samples, with prior meta-learning on images of other diseases only. The results show that DVMN reaches better accuracy on multi-view dataset than simple Matching Network without multi-view handling.


1999 ◽  
Vol 173 ◽  
pp. 289-293 ◽  
Author(s):  
J.R. Donnison ◽  
L.I. Pettit

AbstractA Pareto distribution was used to model the magnitude data for short-period comets up to 1988. It was found using exponential probability plots that the brightness did not vary with period and that the cut-off point previously adopted can be supported statistically. Examination of the diameters of Trans-Neptunian bodies showed that a power law does not adequately fit the limited data available.


VASA ◽  
2014 ◽  
Vol 43 (1) ◽  
pp. 55-61 ◽  
Author(s):  
Konstantinos Tziomalos ◽  
Vasilios Giampatzis ◽  
Stella Bouziana ◽  
Athinodoros Pavlidis ◽  
Marianna Spanou ◽  
...  

Background: Peripheral arterial disease (PAD) is frequently present in patients with acute ischemic stroke. However, there are limited data regarding the association between ankle brachial index (ABI) ≤ 0.90 (which is diagnostic of PAD) or > 1.40 (suggesting calcified arteries) and the severity of stroke and in-hospital outcome in this population. We aimed to evaluate these associations in patients with acute ischemic stroke. Patients and methods: We prospectively studied 342 consecutive patients admitted for acute ischemic stroke (37.4 % males, mean age 78.8 ± 6.4 years). The severity of stroke was assessed with the National Institutes of Health Stroke Scale (NIHSS)and the modified Rankin scale (mRS) at admission. The outcome was assessed with the mRS and dependency (mRS 2 - 5) at discharge and in-hospital mortality. Results: An ABI ≤ 0.90 was present in 24.6 % of the patients whereas 68.1 % had ABI 0.91 - 1.40 and 7.3 % had ABI > 1.40. At admission, the NIHSS score did not differ between the 3 groups (10.4 ± 10.6, 8.3 ± 9.3 and 9.3 ± 9.4, respectively). The mRS score was also comparable in the 3 groups (3.6 ± 1.7, 3.1 ± 1.8 and 3.5 ± 2.3, respectively). At discharge, the mRS score did not differ between the 3 groups (2.9 ± 2.2, 2.3 ± 2.1 and 2.7 ± 2.5, respectively) and dependency rates were also comparable (59.5, 47.6 and 53.3 %, respectively). In-hospital mortality was almost two-times higher in patients with ABI ≤ 0.90 than in patients with ABI 0.91 - 1.40 or > 1.40 but this difference was not significant (10.9, 6.6 and 6.3 %, respectively). Conclusions: An ABI ≤ 0.90 or > 1.40 does not appear to be associated with more severe stroke or worse in-hospital outcome in patients with acute ischemic stroke.


Sign in / Sign up

Export Citation Format

Share Document