scholarly journals Neural Inheritance Relation Guided One-Shot Layer Assignment Search

2020 ◽  
Vol 34 (04) ◽  
pp. 5158-5165
Author(s):  
Rang Meng ◽  
Weijie Chen ◽  
Di Xie ◽  
Yuan Zhang ◽  
Shiliang Pu
Keyword(s):  
Network Performance ◽  
Search Space ◽  
Independent Research ◽  
Neural Architecture ◽  
Network Search ◽  
Layer Assignment ◽  
Search Approach ◽  
The Impact ◽  
First Time ◽  
Strongly Consistent

Layer assignment is seldom picked out as an independent research topic in neural architecture search. In this paper, for the first time, we systematically investigate the impact of different layer assignments to the network performance by building an architecture dataset of layer assignment on CIFAR-100. Through analyzing this dataset, we discover a neural inheritance relation among the networks with different layer assignments, that is, the optimal layer assignments for deeper networks always inherit from those for shallow networks. Inspired by this neural inheritance relation, we propose an efficient one-shot layer assignment search approach via inherited sampling. Specifically, the optimal layer assignment searched in the shallow network can be provided as a strong sampling priori to train and search the deeper ones in supernet, which extremely reduces the network search space. Comprehensive experiments carried out on CIFAR-100 illustrate the efficiency of our proposed method. Our search results are strongly consistent with the optimal ones directly selected from the architecture dataset. To further confirm the generalization of our proposed method, we also conduct experiments on Tiny-ImageNet and ImageNet. Our searched results are remarkably superior to the handcrafted ones under the unchanged computational budgets. The neural inheritance relation discovered in this paper can provide insights to the universal neural architecture search.

scholarly journals MergeNAS: Merge Operations into One for Differentiable Architecture Search

2020 ◽  
Author(s):  
Xiaoxing Wang ◽  
Chao Xue ◽  
Junchi Yan ◽  
Xiaokang Yang ◽  
Yonggang Hu ◽  
...  
Keyword(s):  
Mathematical Formulation ◽  
Search Space ◽  
Search Method ◽  
Test Accuracy ◽  
Search Cost ◽  
Neural Architecture ◽  
Different Types ◽  
Search Approach ◽  
Memory Utilization

Differentiable architecture search (DARTS) has been a promising one-shot architecture search approach for its mathematical formulation and competitive results. However, besides its caused high memory utilization and a large computation requirement, many research works have shown that DARTS also often suffers notable over-fitting and thus does not work robustly for some new tasks. In this paper, we propose a one-shot neural architecture search method referred to as MergeNAS by merging different types of operations e.g. convolutions into one operation. This merge-based approach not only reduces the search cost (about half a GPU day), but also alleviates over-fitting by reducing the redundant parameters. Extensive experiments on different search space and various datasets have been conducted to verify our approach, showing that MergeNAS can converge to a stable architecture and achieve better performance with fewer parameters and search cost. For test accuracy and its stability, MergeNAS outperforms all NAS baseline methods implemented on NAS-Bench-201, including DARTS, ENAS, RS, BOHB, GDAS and hand-crafted ResNet.

scholarly journals Should the impact factor of the year of publication or the last available one be used when evaluating scientists?

2020 ◽  
Vol 18 (3) ◽  
pp. eM01
Author(s):  
Gustavo A. Slafer ◽  
Roxana Savin
Keyword(s):  
Impact Factor ◽  
Impact Factors ◽  
Independent Research ◽  
Common Procedure ◽  
Actual Impact ◽  
The Impact ◽  
First Time ◽  
The Relationship ◽  
Research Centres

Aim of study: A common procedure when evaluating scientists is considering the journal’s quartile of impact factors (within a category), many times considering the quartile in the year of publication instead of the last available ranking. We tested whether the extra work involved in considering the quartiles of each particular year is justifiedArea of study: EuropeMaterial and methods: we retrieved information from all papers published in 2008-2012 by researchers of AGROTECNIO, a centre focused in a range of agri-food subjects. Then, we validated the results observed for AGROTECNIO against five other European independent research centres: Technical University of Madrid (UPM) and the Universities of Nottingham (UK), Copenhagen (Denmark), Helsinki (Finland), and Bologna (Italy).Main results: The relationship between the actual impact of the papers and the impact factor quartile of a journal within its category was not clear, although for evaluations based on recently published papers there might not be much better indicators. We found unnecessary to determine the rank of the journal for the year of publication as the outcome of the evaluation using the last available rank was virtually the same.Research highlights: We confirmed that the journal quality reflects only vaguely the quality of the papers, and reported for the first time evidences that using the journal rank from the particular year that papers were published represents an unnecessary effort and therefore evaluation should be done simply considering the last available rank.

scholarly journals AutoShrink: A Topology-Aware NAS for Discovering Efficient Neural Architecture

2020 ◽  
Vol 34 (04) ◽  
pp. 6829-6836
Author(s):  
Tunhou Zhang ◽  
Hsin-Pai Cheng ◽  
Zhenwen Li ◽  
Feng Yan ◽  
Chengyu Huang ◽  
...  
Keyword(s):  
Cell Structure ◽  
Search Time ◽  
Building Blocks ◽  
Search Space ◽  
Directed Acyclic Graphs ◽  
Neural Architecture ◽  
Cell Structures ◽  
Acyclic Graphs ◽  
Search Approach ◽  
Network Patterns

Resource is an important constraint when deploying Deep Neural Networks (DNNs) on mobile and edge devices. Existing works commonly adopt the cell-based search approach, which limits the flexibility of network patterns in learned cell structures. Moreover, due to the topology-agnostic nature of existing works, including both cell-based and node-based approaches, the search process is time consuming and the performance of found architecture may be sub-optimal. To address these problems, we propose AutoShrink, a topology-aware Neural Architecture Search (NAS) for searching efficient building blocks of neural architectures. Our method is node-based and thus can learn flexible network patterns in cell structures within a topological search space. Directed Acyclic Graphs (DAGs) are used to abstract DNN architectures and progressively optimize the cell structure through edge shrinking. As the search space intrinsically reduces as the edges are progressively shrunk, AutoShrink explores more flexible search space with even less search time. We evaluate AutoShrink on image classification and language tasks by crafting ShrinkCNN and ShrinkRNN models. ShrinkCNN is able to achieve up to 48% parameter reduction and save 34% Multiply-Accumulates (MACs) on ImageNet-1K with comparable accuracy of state-of-the-art (SOTA) models. Specifically, both ShrinkCNN and ShrinkRNN are crafted within 1.5 GPU hours, which is 7.2× and 6.7× faster than the crafting time of SOTA CNN and RNN models, respectively.

Relationship between big five personality model and abusive supervision

2019 ◽  
Vol 12 (2) ◽  
Author(s):  
Bibi Tahira ◽  
Naveed Saif ◽  
Muhammad Haroon ◽  
Sadaqat Ali
Keyword(s):  
Big Five ◽  
Abusive Supervision ◽  
State Universities ◽  
Big Five Personality ◽  
Big Five Model ◽  
The Impact ◽  
First Time ◽  
Supervisor Behavior ◽  
Moderating Role

The current study tries to understand the diverse nature of relationship between personality Big Five Model (PBFM) and student's perception of abusive supervision in higher education institutions of Khyber Pakhtoonkhwa Pakistan. Data was collected in dyads i.e. (supervisors were asked to rate their personality attributes while student were asked to rate the supervisor behavior) through adopted construct. For this purpose, data was collected from three government state universities and one Private Sector University. The focus was on MS/M.Phill and PhD student and their supervisors of the mentioned universities. After measuring normality and validity regression analysis was conducted to assess the impact of supervisor personality characteristics that leads to abusive supervision. Findings indicate interestingly that except agreeableness other four attributes of (PBFM) are play their role for abusive supervision. The results are novel in the nature as for the first time Neuroticism, openness to experience, extraversion and conscientiousness are held responsible for the abusive supervision. The study did not explore the demographic characteristics, and moderating role of organizational culture, justice and interpersonal deviances to understand the strength of relationship in more detail way. Keywords: Personality big five model, abusive supervision, HEIs

The Influence of TRIPS Compliant Patent Laws on Indian Pharmaceutical Industry

2014 ◽  
Vol 4 (1) ◽  
Author(s):  
Rupesh Rastogi ◽  
Virendra Kumar
Keyword(s):  
Pharmaceutical Industry ◽  
Patent Protection ◽  
Uruguay Round ◽  
Pharmaceutical Products ◽  
Trips Agreement ◽  
Patent Law ◽  
Patent Laws ◽  
The Impact ◽  
First Time ◽  
Indian Pharmaceutical Industry

The first legislation in India relating to patents was the Act VI of 1856. The Indian Patents and Design Act, 1911 (Act II of 1911) replaced all the previous Acts. The Act brought patent administration under the management of Controller of Patents for the first time. After Independence, it was felt that the Indian Patents & Designs Act, 1911 was not fulfilling its objective. Various comities were constituted to recommend, framing a patent law which can fulfill the requirement of Indian Industry and people. The Indian Patent Act of 1970 was enacted to achieve the above objectives. The major provisions of the act, provided for process, not the product patents in food, medicines, chemicals with a term of 14 years and 5-7 for chemicals and drugs. The Act enabled Indian citizens to access cheapest medicines in the world and paved a way for exponential growth of Indian Pharmaceutical Industry. TRIPS agreement, which is one of the important results of the Uruguay Round, mandated strong patent protection, especially for pharmaceutical products, thereby allowing the patenting of NCEs, compounds and processes. India is thereby required to meet the minimum standards under the TRIPS Agreement in relation to patents and the pharmaceutical industry. India’s patent legislation must now include provisions for availability of patents for both pharmaceutical products and processes inventions. The present paper examines the impact of change in Indian Patent law on Pharmaceutical Industry.

The impact of traditional plants and their secondary metabolites in the discovery of COVID-19 treatment

2020 ◽  
Vol 26 ◽  
Author(s):  
Shabana Bibi ◽  
Ayesha Sarfraz ◽  
Ghazala Mustafa ◽  
Zeeshan Ahmed ◽  
Muhammad Aurang Zeb ◽  
...  
Keyword(s):  
Secondary Metabolites ◽  
Test System ◽  
Chemical Information ◽  
System Type ◽  
The Family ◽  
Comprehensive Search ◽  
Coronavirus Infection ◽  
The Impact ◽  
First Time ◽  
Plant Sources

Background: Coronavirus Disease-2019 belongs to the family of viruses which cause a serious pneumonia along with fever, breathing issues and infection of lungs for the first time in China and later spread worldwide. Objective: Several studies and clinical trials have been conducted to identify potential drugs and vaccines for Coronavirus Disease-2019. The present study listed natural secondary metabolites identified from plant sources with antiviral properties and could be safer and tolerable treatment for Coronavirus Disease-2019. Methods: A comprehensive search on the reported studies was conducted using different search engine such as Google scholar, SciFinder, Sciencedirect, Medline PubMed, and Scopus for the collection of research articles based on plantderived secondary metabolites, herbal extracts, and traditional medicine for coronavirus infections. Results: Status of COVID-19 worldwide and information of important molecular targets involved in COVID-19 is described and through literature search, is highlighted that numerous plant species and their extracts possess antiviral properties and studied with respect to Coronavirus treatments. Chemical information, plant source, test system type with mechanism of action for each secondary metabolite is also mentioned in this review paper. Conclusion: The present review has listed plants that have presented antiviral potential in the previous coronavirus pandemics and their secondary metabolites which could be significant for the development of novel and a safer drug which could prevent and cure coronavirus infection worldwide.

Individual and Combined Assessment of Ser680Asn FSH Receptor and FSHβ -211 G>T Gene Polymorphisms in Ovarian Response in IVF/ICSI Program

2020 ◽  
Vol 21 ◽  
Author(s):  
Elli Anagnostou ◽  
Alexia Kafkoutsou ◽  
Despina Mavrogianni ◽  
Ekaterini Domali ◽  
Evangelia Dimitroulia ◽  
...  
Keyword(s):  
Gene Polymorphism ◽  
Gene Polymorphisms ◽  
Ovarian Response ◽  
Greek Population ◽  
Control Group ◽  
Genotype Distribution ◽  
Β Subunit ◽  
Molecular Genetic Study ◽  
The Impact ◽  
First Time

Background: Molecular biology tools, such as the detection of single nucleotide polymorphisms (SNPs), have been considered to assist to the management of the ovarian stimulation protocols. Purpose: The aim of this study was to evaluate the impact of two polymorphisms, the Asn680Ser polymorphism of the FSHR gene, and the FSH β subunit (FSHβ) gene polymorphism -211 G>T, in a Greek population of women undergoing IVF/ICSI program in our center. In addition, a control group of fertile women was studied, to verify whether there are differences in the genotype distribution between fertile and infertile population for both polymorphisms, as the FSHβ gene polymorphism -211 G>T is studied for the first time in the Greek population. Results : The FSH β-211 G>T polymorphism, studied for the first time in the Greek infertile population, appears to be quite rare. When studying the two polymorphisms separately, statistically significant differences were obtained that concerned the LH levels. Discussion: According to the combination analysis of the two polymorphisms by the number of alleles, women with 2-3 polymorphic alleles needed more days of stimulation, but there were no differences in pregnancy rates. Conclusion: This molecular genetic study helps to elucidate whether the polygenic combination of the Asn680Ser and FSH β subunit -211 G>T gene polymorphisms is of additive value in the prediction of ovarian response to exogenous gonadotropins.

Thermal-aware Output Polarity Selection Based on And-Inverter Graph Manipulation

2019 ◽  
Vol 12 (1) ◽  
pp. 30-39 ◽  
Author(s):  
Apangshu Das ◽  
Sambhu Nath Pradhan
Keyword(s):  
Integrated Circuits ◽  
Power Density ◽  
Logic Synthesis ◽  
Programmable Logic Arrays ◽  
Power Switching ◽  
Nsga Ii ◽  
Multi Level ◽  
Search Approach ◽  
Efficient Power ◽  
The Impact

Background: Output polarity of the sub-function is generally considered to reduce the area and power of a circuit at the two-level realization. Along with area and power, the power-density is also one of the significant parameter which needs to be consider, because power-density directly converges to circuit temperature. More than 50% of the modern day integrated circuits are damaged due to excessive overheating. Methods: This work demonstrates the impact of efficient power density based logic synthesis (in the form of suitable polarity selection of sub-function of Programmable Logic Arrays (PLAs) for its multilevel realization) for the reduction of temperature. Two-level PLA optimization using output polarity selection is considered first and compared with other existing techniques and then And-Invert Graphs (AIG) based multi-level realization has been considered to overcome the redundant solution generated in two-level synthesis. AIG nodes and associated power dissipation can be reduced by rewriting, refactoring and balancing technique. Reduction of nodes leads to the reduction of the area but on the contrary increases power and power density of the circuit. A meta-heuristic search approach i.e., Nondominated Sorting Genetic Algorithm-II (NSGA-II) is proposed to select the suitable output polarity of PLA sub-functions for its optimal realization. Results: Best power density based solution saves up to 8.29% power density compared to ‘espresso – dopo’ based solutions. Around 9.57% saving in area and 9.67% saving in power (switching activity) are obtained with respect to ‘espresso’ based solution using NSGA-II. Conclusion: Suitable output polarity realized circuit is converted into multi-level AIG structure and synthesized to overcome the redundant solution at the two-level circuit. It is observed that with the increase in power density, the temperature of a particular circuit is also increases.

Reconstituting the International, 1918–1923

2017 ◽  
Author(s):  
Talbot C. Imlay
Keyword(s):  
Seismic Events ◽  
Great War ◽  
Bolshevik Revolution ◽  
Meaning And Purpose ◽  
Post War ◽  
The Impact ◽  
First Time ◽  
The Great War ◽  
Do So ◽  
Lengthy Process

This chapter examines the post-war efforts of European socialists to reconstitute the Socialist International. Initial efforts to cooperate culminated in an international socialist conference in Berne in February 1919 at which socialists from the two wartime camps met for the first time. In the end, however, it would take four years to reconstitute the International with the creation of the Labour and Socialist International (LSI) in 1923. That it took so long to do so is a testimony to the impact of the Great War and to the Bolshevik revolution. Together, these two seismic events compelled socialists to reconsider the meaning and purpose of socialism. The search for answers sparked prolonged debates between and within the major parties, profoundly reconfiguring the pre-war world of European socialism. One prominent stake in this lengthy process, moreover, was the nature of socialist internationalism—both its content and its functioning.

scholarly journals HTPC: heterogeneous traffic-aware partition coding for random packet spraying in data center networks

2021 ◽  
Vol 10 (1) ◽  
Author(s):  
Jiawei Huang ◽  
Shiqi Wang ◽  
Shuping Li ◽  
Shaojun Zou ◽  
Jinbin Hu ◽  
...  
Keyword(s):  
Data Center ◽  
Large Scale ◽  
Network Performance ◽  
Rooted Tree ◽  
Heterogeneous Traffic ◽  
Data Center Networks ◽  
Packet Reordering ◽  
Traffic Characteristics ◽  
Network Utilization ◽  
The Impact

AbstractModern data center networks typically adopt multi-rooted tree topologies such leaf-spine and fat-tree to provide high bisection bandwidth. Load balancing is critical to achieve low latency and high throughput. Although the per-packet schemes such as Random Packet Spraying (RPS) can achieve high network utilization and near-optimal tail latency in symmetric topologies, they are prone to cause significant packet reordering and degrade the network performance. Moreover, some coding-based schemes are proposed to alleviate the problem of packet reordering and loss. Unfortunately, these schemes ignore the traffic characteristics of data center network and cannot achieve good network performance. In this paper, we propose a Heterogeneous Traffic-aware Partition Coding named HTPC to eliminate the impact of packet reordering and improve the performance of short and long flows. HTPC smoothly adjusts the number of redundant packets based on the multi-path congestion information and the traffic characteristics so that the tailing probability of short flows and the timeout probability of long flows can be reduced. Through a series of large-scale NS2 simulations, we demonstrate that HTPC reduces average flow completion time by up to 60% compared with the state-of-the-art mechanisms.

Sign in / Sign up

Export Citation Format

Share Document