Review on biomass feedstocks, pyrolysis mechanism and physicochemical properties of biochar: State-of-the-art framework to speed up vision of circular bioeconomy

Connected components is a fundamental kernel in graph applications. The fastest existing multicore algorithms for solving graph connectivity are based on some form of edge sampling and/or linking and compressing trees. However, many combinations of these design choices have been left unexplored. In this paper, we design the ConnectIt framework, which provides different sampling strategies as well as various tree linking and compression schemes. ConnectIt enables us to obtain several hundred new variants of connectivity algorithms, most of which extend to computing spanning forest. In addition to static graphs, we also extend ConnectIt to support mixes of insertions and connectivity queries in the concurrent setting. We present an experimental evaluation of ConnectIt on a 72-core machine, which we believe is the most comprehensive evaluation of parallel connectivity algorithms to date. Compared to a collection of state-of-the-art static multicore algorithms, we obtain an average speedup of 12.4x (2.36x average speedup over the fastest existing implementation for each graph). Using ConnectIt, we are able to compute connectivity on the largest publicly-available graph (with over 3.5 billion vertices and 128 billion edges) in under 10 seconds using a 72-core machine, providing a 3.1x speedup over the fastest existing connectivity result for this graph, in any computational setting. For our incremental algorithms, we show that our algorithms can ingest graph updates at up to several billion edges per second. To guide the user in selecting the best variants in ConnectIt for different situations, we provide a detailed analysis of the different strategies. Finally, we show how the techniques in ConnectIt can be used to speed up two important graph applications: approximate minimum spanning forest and SCAN clustering.

Download Full-text

Data-Efficient Sensor Upgrade Path Using Knowledge Distillation

Sensors ◽

10.3390/s21196523 ◽

2021 ◽

Vol 21 (19) ◽

pp. 6523

Author(s):

Pieter Van Van Molle ◽

Cedric De De Boom ◽

Tim Verbelen ◽

Bert Vankeirsbilck ◽

Jonas De De Vylder ◽

...

Keyword(s):

Deep Neural Networks ◽

State Of The Art ◽

Original Data ◽

Radar Data ◽

Teacher Supervision ◽

Multispectral Images ◽

Test Set ◽

Time To Market ◽

Speed Up ◽

Knowledge Distillation

Deep neural networks have achieved state-of-the-art performance in image classification. Due to this success, deep learning is now also being applied to other data modalities such as multispectral images, lidar and radar data. However, successfully training a deep neural network requires a large reddataset. Therefore, transitioning to a new sensor modality (e.g., from regular camera images to multispectral camera images) might result in a drop in performance, due to the limited availability of data in the new modality. This might hinder the adoption rate and time to market for new sensor technologies. In this paper, we present an approach to leverage the knowledge of a teacher network, that was trained using the original data modality, to improve the performance of a student network on a new data modality: a technique known in literature as knowledge distillation. By applying knowledge distillation to the problem of sensor transition, we can greatly speed up this process. We validate this approach using a multimodal version of the MNIST dataset. Especially when little data is available in the new modality (i.e., 10 images), training with additional teacher supervision results in increased performance, with the student network scoring a test set accuracy of 0.77, compared to an accuracy of 0.37 for the baseline. We also explore two extensions to the default method of knowledge distillation, which we evaluate on a multimodal version of the CIFAR-10 dataset: an annealing scheme for the hyperparameter α and selective knowledge distillation. Of these two, the first yields the best results. Choosing the optimal annealing scheme results in an increase in test set accuracy of 6%. Finally, we apply our method to the real-world use case of skin lesion classification.

Download Full-text

A Fast Link Initialization Protocol for Beam-Steering based Cellular Backhaul Systems

Network Protocols and Algorithms ◽

10.5296/npa.v7i2.7440 ◽

2015 ◽

Vol 7 (2) ◽

pp. 113

Author(s):

Markus Petri ◽

Marcus Ehrig ◽

Markus Günther

Keyword(s):

State Of The Art ◽

Beam Steering ◽

Small Cell ◽

Base Stations ◽

Data Traffic ◽

Mobile Data ◽

User Expectations ◽

Network Topologies ◽

Speed Up ◽

Directive Antennas

<p>To deal with the enormous increase of mobile data traffic, new cellular network topologies are necessary. The reduction of cell area and the usage of light-weighted base stations serving only a handful of users, commonly known as the small cell approach, seems to be a suitable solution addressing changes in user expectations and usage scenarios. This paper is an extended version of [1], where current challenges of small cell deployments were presented from a backhaul perspective. A mesh-type backhaul network topology based on beam-steering millimeter-wave systems was proposed as a future-proof solution. In this paper, we focus on a link initialization protocol for beam-steering with highly directive antennas. Special requirements and problems for link setup are analyzed. Based on that, a fast protocol for link initialization is presented and it is evaluated in terms of the resulting initialization speed-up compared to state-of-the-art solutions. Furthermore, a potential approach for extending the fast link initialization protocol to support point-to-multipoint connections is given.</p>

Download Full-text

Text Mining in Cybersecurity: Exploring Threats and Opportunities

Multimodal Technologies and Interaction ◽

10.3390/mti3030062 ◽

2019 ◽

Vol 3 (3) ◽

pp. 62 ◽

Cited By ~ 1

Author(s):

Maaike H. T. de Boer ◽

Babette J. Bakker ◽

Erik Boertjes ◽

Mike Wilmer ◽

Stephan Raaijmakers ◽

...

Keyword(s):

Information Retrieval ◽

Text Mining ◽

State Of The Art ◽

Data Sources ◽

Added Value ◽

Design Decision ◽

User Evaluation ◽

User Query ◽

Overall Evaluation ◽

Speed Up

The number of cyberattacks on organizations is growing. To increase cyber resilience, organizations need to obtain foresight to anticipate cybersecurity vulnerabilities, developments, and potential threats. This paper describes a tool that combines state of the art text mining and information retrieval techniques to explore the opportunities of using these techniques in the cybersecurity domain. Our tool, the Horizon Scanner, can scrape and store data from websites, blogs and PDF articles, and search a database based on a user query, show textual entities in a graph, and provide and visualize potential trends. The aim of the Horizon Scanner is to help experts explore relevant data sources for potential threats and trends and to speed up the process of foresight. In a requirements session and user evaluation of the tool with cyber experts from the Dutch Defense Cyber Command, we explored whether the Horizon Scanner tool has the potential to fulfill its aim in the cybersecurity domain. Although the overall evaluation of the tool was not as good as expected, some aspects of the tool were found to have added value, providing us with valuable insights into how to design decision support for forecasting analysts.

Download Full-text

SubMito-PSPCP: Predicting Protein Submitochondrial Locations by Hybridizing Positional Specific Physicochemical Properties with Pseudoamino Acid Compositions

BioMed Research International ◽

10.1155/2013/263829 ◽

2013 ◽

Vol 2013 ◽

pp. 1-7 ◽

Cited By ~ 13

Author(s):

Pufeng Du ◽

Yuan Yu

Keyword(s):

Physicochemical Properties ◽

Outer Membrane ◽

Acid Composition ◽

State Of The Art ◽

Mitochondrial Protein ◽

Protein Sequences ◽

Online Service ◽

Inner Membrane ◽

Future Studies ◽

General Method

Knowing the submitochondrial location of a mitochondrial protein is an important step in understanding its function. We developed a new method for predicting protein submitochondrial locations by introducing a new concept: positional specific physicochemical properties. With the framework of general form pseudoamino acid compositions, our method used only about 100 features to represent protein sequences, which is much simpler than the existing methods. On the dataset of SubMito, our method achieved over 93% overall accuracy, with 98.60% for inner membrane, 93.90% for matrix, and 70.70% for outer membrane, which are comparable to all state-of-the-art methods. As our method can be used as a general method to upgrade all pseudoamino-acid-composition-based methods, it should be very useful in future studies. We implement our method as an online service: SubMito-PSPCP.

Download Full-text

Truss Decomposition using Triangle Graphs

10.21203/rs.3.rs-819379/v1 ◽

2021 ◽

Author(s):

Mohsen Rezvani ◽

Mojtaba Rezvani

Keyword(s):

Social Networks ◽

Large Scale ◽

State Of The Art ◽

Decomposition Algorithms ◽

Scalable Algorithm ◽

Large Scale Network ◽

Triangle Graph ◽

Speed Up ◽

Scale Network ◽

High Degree

Abstract Recent studies have shown that social networks exhibit interesting characteristics such as community structures, i.e., vertexes can be clustered into communities that are densely connected together and loosely connected to other vertices. In order to identify communities, several definitions have been proposed that can characterize the density of connections among vertices in the networks. Dense triangle cores, also known as $k$-trusses, are subgraphs in which every edge participates at least $k-2$ triangles (a clique of size 3), exhibiting a high degree of cohesiveness among vertices. There are a number of research works that propose $k$-truss decomposition algorithms. However, existing in-memory algorithms for computing $k$-truss are inefficient for handling today’s massive networks. In this paper, we propose an efficient, yet scalable algorithm for finding $k$-trusses in a large-scale network. To this end, we propose a new structure, called triangle graph to speed up the process of finding the $k$-trusses and prove the correctness and efficiency of our method. We also evaluate the performance of the proposed algorithms through extensive experiments using real-world networks. The results of comprehensive experiments show that the proposed algorithms outperform the state-of-the-art methods by several orders of magnitudes in running time.

Download Full-text

Chemformer: A Pre-Trained Transformer for Computational Chemistry

Machine Learning: Science and Technology ◽

10.1088/2632-2153/ac3ffb ◽

2021 ◽

Author(s):

Ross Irwin ◽

Spyridon Dimitriadis ◽

Jiazhen He ◽

Esben Jannik Bjerrum

Keyword(s):

Computational Chemistry ◽

State Of The Art ◽

Direct Synthesis ◽

Single Application ◽

Improve Performance ◽

Molecular Line ◽

Benchmark Datasets ◽

Speed Up

Abstract Transformer models coupled with Simplified Molecular Line Entry System (SMILES) have recently proven to be a powerful combination for solving challenges in cheminformatics. These models, however, are often developed specifically for a single application and can be very resource-intensive to train. In this work we present Chemformer model – a Transformerbased model which can be quickly applied to both sequence-to-sequence and discriminative cheminformatics tasks. Additionally, we show that self-supervised pre-training can improve performance and significantly speed up convergence on downstream tasks. On direct synthesis and retrosynthesis prediction benchmark datasets we publish state-of-the-art results for top- 1 accuracy. We also improve on existing approaches for a molecular optimisation task and show that Chemformer can optimise on multiple discriminative tasks simultaneously. Models, datasets and code will be made available after publication.

Download Full-text

Stacking models for nearly optimal link prediction in complex networks

Proceedings of the National Academy of Sciences ◽

10.1073/pnas.1914950117 ◽

2020 ◽

Vol 117 (38) ◽

pp. 23393-23400 ◽

Cited By ~ 4

Author(s):

Amir Ghasemian ◽

Homa Hosseinmardi ◽

Aram Galstyan ◽

Edoardo M. Airoldi ◽

Aaron Clauset

Keyword(s):

Social Networks ◽

Data Collection ◽

Real World ◽

Link Prediction ◽

State Of The Art ◽

Network Data ◽

Prediction Errors ◽

Partially Observed ◽

Speed Up ◽

Large Corpus

Most real-world networks are incompletely observed. Algorithms that can accurately predict which links are missing can dramatically speed up network data collection and improve network model validation. Many algorithms now exist for predicting missing links, given a partially observed network, but it has remained unknown whether a single best predictor exists, how link predictability varies across methods and networks from different domains, and how close to optimality current methods are. We answer these questions by systematically evaluating 203 individual link predictor algorithms, representing three popular families of methods, applied to a large corpus of 550 structurally diverse networks from six scientific domains. We first show that individual algorithms exhibit a broad diversity of prediction errors, such that no one predictor or family is best, or worst, across all realistic inputs. We then exploit this diversity using network-based metalearning to construct a series of “stacked” models that combine predictors into a single algorithm. Applied to a broad range of synthetic networks, for which we may analytically calculate optimal performance, these stacked models achieve optimal or nearly optimal levels of accuracy. Applied to real-world networks, stacked models are superior, but their accuracy varies strongly by domain, suggesting that link prediction may be fundamentally easier in social networks than in biological or technological networks. These results indicate that the state of the art for link prediction comes from combining individual algorithms, which can achieve nearly optimal predictions. We close with a brief discussion of limitations and opportunities for further improvements.

Download Full-text

A Multibranch Object Detection Method for Traffic Scenes

Computational Intelligence and Neuroscience ◽

10.1155/2019/3679203 ◽

2019 ◽

Vol 2019 ◽

pp. 1-16

Author(s):

Jiangfan Feng ◽

Fanjie Wang ◽

Siqin Feng ◽

Yongrong Peng

Keyword(s):

Neural Network ◽

Object Detection ◽

Convolutional Neural Network ◽

Detection Method ◽

State Of The Art ◽

Recall Rate ◽

Small Scale ◽

Feature Maps ◽

Time Requirements ◽

Speed Up

The performance of convolutional neural network- (CNN-) based object detection has achieved incredible success. Howbeit, existing CNN-based algorithms suffer from a problem that small-scale objects are difficult to detect because it may have lost its response when the feature map has reached a certain depth, and it is common that the scale of objects (such as cars, buses, and pedestrians) contained in traffic images and videos varies greatly. In this paper, we present a 32-layer multibranch convolutional neural network named MBNet for fast detecting objects in traffic scenes. Our model utilizes three detection branches, in which feature maps with a size of 16 × 16, 32 × 32, and 64 × 64 are used, respectively, to optimize the detection for large-, medium-, and small-scale objects. By means of a multitask loss function, our model can be trained end-to-end. The experimental results show that our model achieves state-of-the-art performance in terms of precision and recall rate, and the detection speed (up to 33 fps) is fast, which can meet the real-time requirements of industry.

Download Full-text

A New Generalized Projection and Its Application to Acceleration of Audio Declipping

Axioms ◽

10.3390/axioms8030105 ◽

2019 ◽

Vol 8 (3) ◽

pp. 105

Author(s):

Pavel Rajmic ◽

Pavel Záviška ◽

Vítězslav Veselý ◽

Ondřej Mokrý

Keyword(s):

Signal Processing ◽

Convex Optimization ◽

Linear Operator ◽

Explicit Formula ◽

Convex Sets ◽

State Of The Art ◽

Computational Cost ◽

The Other ◽

Speed Up ◽

Low Computational Cost

In convex optimization, it is often inevitable to work with projectors onto convex sets composed with a linear operator. Such a need arises from both the theory and applications, with signal processing being a prominent and broad field where convex optimization has been used recently. In this article, a novel projector is presented, which generalizes previous results in that it admits to work with a broader family of linear transforms when compared with the state of the art but, on the other hand, it is limited to box-type convex sets in the transformed domain. The new projector is described by an explicit formula, which makes it simple to implement and requires a low computational cost. The projector is interpreted within the framework of the so-called proximal splitting theory. The convenience of the new projector is demonstrated on an example from signal processing, where it was possible to speed up the convergence of a signal declipping algorithm by a factor of more than two.

Download Full-text