scholarly journals Learning with Selective Forgetting

2021 ◽  
Author(s):  
Takashi Shibata ◽  
Go Irie ◽  
Daiki Ikami ◽  
Yu Mitsuzumi
Keyword(s):  
Lifelong Learning ◽  
Original Data ◽  
Ethical Considerations ◽  
The Past ◽  
Synthetic Signal ◽  
Training Samples ◽  
Benchmark Datasets ◽  
Privacy Issues ◽  
New Framework

Lifelong learning aims to train a highly expressive model for a new task while retaining all knowledge for previous tasks. However, many practical scenarios do not always require the system to remember all of the past knowledge. Instead, ethical considerations call for selective and proactive forgetting of undesirable knowledge in order to prevent privacy issues and data leakage. In this paper, we propose a new framework for lifelong learning, called Learning with Selective Forgetting, which is to update a model for the new task with forgetting only the selected classes of the previous tasks while maintaining the rest. The key is to introduce a class-specific synthetic signal called mnemonic code. The codes are "watermarked" on all the training samples of the corresponding classes when the model is updated for a new task. This enables us to forget arbitrary classes later by only using the mnemonic codes without using the original data. Experiments on common benchmark datasets demonstrate the remarkable superiority of the proposed method over several existing methods.

Segmentation of Tissue-Injured Melanoma Convolution Neural Networks

2021 ◽  
Vol 18 (4) ◽  
pp. 1256-1262
Author(s):  
C. Hemalatha ◽  
S. Satheesh ◽  
N. Kamal ◽  
C. Devi ◽  
A. Vinothkumar ◽  
...  
Keyword(s):  
Neural Networks ◽  
Skin Lesions ◽  
Convolution Neural Networks ◽  
Skin Imaging ◽  
The Past ◽  
Training Samples ◽  
Basic Work ◽  
Benchmark Datasets ◽  
Variable Success ◽  
Ann Classification

In global dermatological conditions, skin lesions are significant. Curable early in the diagnosis, only skin lesions can be accurately identified by highly trained dermatologists. Around 21 million patients are diagnosed with this disease and more than 10.12 million deaths worldwide. This paper presents basic work for the detection and ensuing purpose of the CNN to dermoscopic images of skin lesions with cancerous inclination. The models proposed are trained and evaluated in the 2018 International Skin Imaging Collaboration challenge, comprising 2100 training samples and 750 test samples, on normal benchmark datasets. Skin-injured images were mainly segment based on person thresholds for channel intensity. The images were added to CNN to extract features. The extracted characteristics were then used to classify the associated ANN classification. In the past, many approaches have been used to diagnose subjects with variable success levels. The methodology described in this paper showed associated accuracy of 97.13% in comparison to the previous best of ninety seven.

Recruitment Policies of Non-EU/EEA Nurses and Ethical Issues

2018 ◽  
Author(s):  
Angèle Flora Mendy
Keyword(s):  
Public Opinion ◽  
Health Professionals ◽  
Ethical Issues ◽  
Health Workers ◽  
Ethical Considerations ◽  
Host Countries ◽  
Ethical Argument ◽  
The Past ◽  
The United Kingdom ◽  
Institutional Legacies

By examining policies of recruiting non-EU/EEA health workers and how ethical considerations are taken into account when employing non-EU/EEA nurses in the United Kingdom, France, and Switzerland, this chapter intends to show that the use of the so-called ‘ethical’ argument to convince national public opinion of the relevance of restrictive recruitment policies is recent (since the 1990s). The analysis highlights the fact that in addition to the institutional legacies, qualification and skills—through the process of their recognition—play an important role in the opening or restriction of the labour market to health professionals from the Global South. The legacy of the past also largely determines the place offered to non-EU/EEA health professionals in the different health systems of host countries.

Beyond Deep Learning: An Econometric Example

2020 ◽  
Vol 28 (Supp01) ◽  
pp. 31-38 ◽  
Author(s):  
Ruofan Liao ◽  
Paravee Maneejuk ◽  
Songsak Sriboonchitta
Keyword(s):  
Neural Networks ◽  
Deep Learning ◽  
Prediction Models ◽  
Original Data ◽  
Parametric Model ◽  
Parametric Models ◽  
The Past ◽  
Currency Exchange ◽  
Currency Exchange Rate ◽  
Linear And Nonlinear

In the past, in many areas, the best prediction models were linear and nonlinear parametric models. In the last decade, in many application areas, deep learning has shown to lead to more accurate predictions than the parametric models. Deep learning-based predictions are reasonably accurate, but not perfect. How can we achieve better accuracy? To achieve this objective, we propose to combine neural networks with parametric model: namely, to train neural networks not on the original data, but on the differences between the actual data and the predictions of the parametric model. On the example of predicting currency exchange rate, we show that this idea indeed leads to more accurate predictions.

Egyptology and Global History: Between Geocultural Power and the Crisis of Humanities

2021 ◽  
Vol 13 (1-2) ◽  
pp. 29-76
Author(s):  
Juan Carlos Moreno García
Keyword(s):  
Social Sciences ◽  
Global History ◽  
Historical Perspectives ◽  
The Past ◽  
Gradual Loss ◽  
Dynamic Flows ◽  
Western Hegemony ◽  
Intellectual Environment ◽  
New Strategies ◽  
New Framework

Abstract Globalization, the decline of Western hegemony, and the rise of new political and economic actors, particularly in East Asia, are concomitant with the emergence of more encompassing historical perspectives, attentive to the achievements and historical trajectories of other regions of the world. Global history provides thus a new framework to understanding our past that challenges former views based on the cultural needs, values, and expectations of the West. This means that humanities and social sciences are subject to intense scrutiny and pressed to adapt themselves to a changing cultural, academic, and intellectual environment. However, this process is hindered by the gradual loss of their former prestige and by the increasing influence of economics in the reorganization of the educational, research, and cultural agenda according to market-oriented criteria. The result is that the mobilization of the past increasingly conforms to new strategies in which connectivity, trading, and diplomatic interests, as well as integration in dynamic flows of wealth, appear of paramount importance. Egyptology is not alien to these challenges, which will in all probability reshape its very foundations in the foreseeable future.

scholarly journals Event conjunction: How the hippocampus integrates episodic memories across event boundaries

2019 ◽  
Author(s):  
Benjamin James Griffiths ◽  
Lluís Fuentemilla
Keyword(s):  
Related Event ◽  
Gamma Model ◽  
Apparent Contradiction ◽  
The Past ◽  
Continuous Stream ◽  
Event Model ◽  
Episodic Memories ◽  
Hippocampal Theta ◽  
Conceptual Link ◽  
New Framework

Our lives are a continuous stream of experience. Our episodic memories, however, have a definitive beginning, middle and end. Theories of event segmentation suggest that salient changes in our environment produce event boundaries which partition the past from the present and, as a result, produce discretised memories. However, event boundaries cannot completely discretise two memories; any shared conceptual link will eagerly integrate these memories. Here, we present a new framework inspired by electrophysiological research that resolves this apparent contradiction. At its heart, the framework proposes that hippocampal theta-gamma coupling maintains a highly abstract model of an ongoing event and serves to encode this model as an episodic memory. When a second but related event begins, this theta-gamma model is rapidly reconstructed within the hippocampus where new details of the second event can be appended to the existing event model. The event conjunction framework is the first electrophysiological explanation of how event memories can be formed at, and integrated across, event boundaries.

scholarly journals Finding the Number of Clusters in Data and Better Initial Centers for K-means Algorithm

2020 ◽  
Vol 12 (6) ◽  
pp. 1-20
Author(s):  
Ahmed Fahim ◽  
Keyword(s):  
Data Clustering ◽  
Linear Time ◽  
Original Data ◽  
Local Minima ◽  
Expected Number ◽  
Open Problems ◽  
Number Of Clusters ◽  
Benchmark Datasets ◽  
Selection Of

The k-means is the most well-known algorithm for data clustering in data mining. Its simplicity and speed of convergence to local minima are the most important advantages of it, in addition to its linear time complexity. The most important open problems in this algorithm are the selection of initial centers and the determination of the exact number of clusters in advance. This paper proposes a solution for these two problems together; by adding a preprocess step to get the expected number of clusters in data and better initial centers. There are many researches to solve each of these problems separately, but there is no research to solve both problems together. The preprocess step requires o(n log n); where n is size of the dataset. This preprocess step aims to get initial portioning of data without determining the number of clusters in advance, then computes the means of initial clusters. After that we apply k-means on original data using the resulting information from the preprocess step to get the final clusters. We use many benchmark datasets to test the proposed method. The experimental results show the efficiency of the proposed method.

scholarly journals The Challenges of Effectively Anonymizing Network Data

2013 ◽  
pp. 15-22
Author(s):  
Ch. Himabindu
Keyword(s):  
Network Security ◽  
Information Security ◽  
Data Collection ◽  
Network Data ◽  
Sensitive Information ◽  
Security Testing ◽  
The Past ◽  
Technical Leadership ◽  
Security Research ◽  
Privacy Issues

The availability of realistic network data plays a significant role in fostering collaboration and ensuring U.S. technical leadership in network security research. Unfortunately, a host of technical, legal, policy, and privacy issues limit the ability of operators to produce datasets for information security testing. In an effort to help overcome these limitations, several data collection efforts (e.g., CRAWDAD[14], PREDICT [34]) have been established in the past few years. The key principle used in all of these efforts to assure low-risk, high-value data is that of trace anonymization—the process of sanitizing data before release so that potentially sensitive information cannot be extracted.

scholarly journals Adaptive reuse process of the Historic Urban Landscape post-Covid-19. The potential of the inner areas for a “new normal”

2020 ◽  
Vol 5 (2) ◽  
pp. 87
Author(s):  
Maria Rita Pinto ◽  
Serena Viola ◽  
Katia Fabbricatti ◽  
Maria Giovanna Pacifico
Keyword(s):  
Urban Landscape ◽  
Adaptive Reuse ◽  
Health And Safety ◽  
Security Requirements ◽  
Health Security ◽  
Settlement Systems ◽  
The Past ◽  
The Face ◽  
The City ◽  
New Framework

<p class="Abstracttext-VITRUVIOCxSpFirst">Often in the past, the great disasters (environmental calamities, earthquakes, epidemics) activated unexpressed energies, triggering transformations of the built environment, able to give rise unexpected conditions of economic, cultural and social development. The fragility of settlement systems in the face of unexpected threats brings out the need for a new planning, changing our gaze on the city.</p><p class="Abstracttext-VITRUVIOCxSpMiddle">The new framework of needs drawn by the pandemic and the renewed sensitivity towards the combination of health – sustainability, rekindle the spotlight on inner areas. These emerged as "reservoirs of resilience", areas to look at, in order to reach an eco-systemic balance.</p><p class="Abstracttext-VITRUVIOCxSpMiddle">The aim of the paper is to return an experience of adaptive reuse of the Historical Urban Landscape in an inner area of Southern Italy, where the needs of health and safety of the community are integrated with the transmission of the built heritage to future generations. The goal is the promotion of inclusive prosperity scenarios, towards the so-called "new normality".</p><p class="Abstracttext-VITRUVIOCxSpLast">Starting from an in-depth literature review on the cases of pandemics in history and the strategies implemented, the research identifies health security requirements at the scale of the Historical Urban Landscape and design solutions aimed at reactivating lost synergies between communities and places.</p>

Foster employability and fight social exclusion through the development of lifelong learning (LLL) key-competences: reviewing twenty years of LLL policies

2021 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Andrea Ceschi ◽  
Marco Perini ◽  
Andrea Scalco ◽  
Monica Pentassuglia ◽  
Elisa Righetti ◽  
...  
Keyword(s):  
Lifelong Learning ◽  
Social Exclusion ◽  
Vocational Training ◽  
Design Methodology ◽  
Decision Processes ◽  
Content Type ◽  
The Past ◽  
New Methods ◽  
E Learning ◽  
Key Competences

Purpose This study aims to provide an overview of the past two decades of lifelong learning (LLL) policies for enhancing employability and reduce social exclusion in young people of European countries through the development of the so-called LLL key-competences. Design/methodology/approach Built on a quasi-systematic review, this contribution explores traditional and new methods for promoting the LLL transition, and then employability, in young adults (e.g. apprenticeship, vocational training, e-learning, etc.). Findings It argues the need to identify all the possible approaches able to support policymakers, as they can differently impact key-competence development. Originality/value Finally, based on the consolidated EU policy experience, we propose a strategy of implementation of the LLL programmes that facilitates the institutions’ decision processes for policy-making through the use of decisional support system.

An Overview of Learning and Teaching From the Past to the Present

2020 ◽  
pp. 1-19
Author(s):  
Cihad Şentürk ◽  
Gökhan Baş
Keyword(s):  
Lifelong Learning ◽  
Learning By Doing ◽  
Learning And Teaching ◽  
The Past ◽  
The World ◽  
Technological Developments ◽  
External Stimuli

Just like any other area in the world, which is quickly changing and converting in line with the scientific and technological developments, the models, approaches, and paradigms set forth as elements of learning and teaching have also undergone alterations and transformations from past to present. While the learning-teaching theories and approaches in the last century, which are based on perennialist and essentialist education philosophies and positivism paradigm, were deeming the learners as passive receivers of external stimuli and focused on the observable and measurable behaviors, the learning-teaching theories and approaches in our century, which are developed around the progressivism and re-constructionism philosophies and post-positivism paradigm, have an understanding that allocates the responsibility to the learner and adopts a lifelong learning by doing and experiencing. In this chapter, a general outlook on the learning and teaching theories and approaches will be briefly carried out.

Sign in / Sign up

Export Citation Format

Share Document