Given a simplicial complex with n simplices, we consider the Connected Subsurface Recognition (c-SR) problem of finding a subcomplex that is homeomorphic to a given connected surface with a fixed boundary. We also stu...
详细信息
Wireless Sensor Systems (WSN) is a broad, exciting area with new perspectives and growing growth over the past decades, where more research is being done. WSNs contain many (hundreds of thousands) of micro-sized, chea...
Wireless Sensor Systems (WSN) is a broad, exciting area with new perspectives and growing growth over the past decades, where more research is being done. WSNs contain many (hundreds of thousands) of micro-sized, cheap chips, powered by low-cost wireless interconnected batteries. These chips are called nodes, which could be of several types, including acoustic, radar, low-frequency magnetic, thermal, and visual sampling frequencies. Many applications are based nowadays on WSN, such as environmental control, smart cities, wildlife monitoring, vehicles and infrastructure, natural disasters, home security, underwater investigations, military, airplane surveillance and drone body sensors. One of the aims of the paper is to develop a smart architecture based on intelligent microprocessor sensors, WSN802GCA/GPA and the software needed to operate with a TINI control unit. TINI allows this system to be connected to Ethernet via TCP/I$\mathbf{P}$
We consider geometric problems on planar n2-point sets in the congested clique model. Initially, each node in the n-clique network holds a batch of n distinct points in the Euclidean plane given by O(log n)-bit coordi...
详细信息
Consensus achievement is a scalable process, which relies on the mutual 'Proof of something' among the consenting parties. Not every application has the same consensus needs: a classification of escalating com...
详细信息
Several domains increasingly rely on machine learning in their applications. The resulting heavy dependence on data has led to the emergence of various laws and regulations around data ethics and privacy and growing a...
详细信息
The thesis data is increasing yearly; classification is one method that can be used to obtain added information from this case. The classification that suits this case is the multi-label classification. The multi-labe...
详细信息
The thesis data is increasing yearly; classification is one method that can be used to obtain added information from this case. The classification that suits this case is the multi-label classification. The multi-label classification with the Problem Transformation approach is a flexible approach. In this study, a comparison of the Problem Transformation methods, namely Binary Relevance, Label Powerset, and Classifier Chain, was carried out with the Multinomial Naïve Bayes algorithm as an estimator. As a result, the Classifier Chain method gives results that tend to be better, with the value of Hamming Loss being 0.0321 and an accuracy of 64.60%.
Rank and select queries are the fundamental building blocks of the compressed data structures. On a given bit string of length $n$ , counting the number of set bits up to a certain position is named as the rank, and f...
Rank and select queries are the fundamental building blocks of the compressed data structures. On a given bit string of length $n$ , counting the number of set bits up to a certain position is named as the rank, and finding the position of the $k$ th set bit is the select query. We present a new data structure and the procedures on it to support rank/select operations. The proposed scheme introduces ( $\frac{\log 2m}{d}+\frac{\log n}{s\cdot d}$ ) overhead bits per each bit over the $n$ -bits long input bit string, where $d$ is the inner-block size in bits, $s$ is the number of inner-blocks in a super-block, and $m$ is a properly chosen constant modulus value. When compared to the previous two-level hierarchical data structures that generate $(\frac{\log(s\cdot d)}{d}+\frac{\log n}{s\cdot d})$ overhead bits per bit, the new approach reduces the space consumption significantly with proper selection of the parameters. With the new data structure, the rank queries are usually (≈ 90% of the time) executed in $O(t_{d})$ time, where $O(t_{d})$ is the time required to compute a rank in an inner-block of length $d$ -bits, which is assumed to be constant via the wide-register instructions in modern processors. Seldom, it may require to investigate more than one block, where on average this is observed to be around two blocks, empirically. We provide probabilistic analyses on how to choose the appropriate parameters and present several trade-offs to guarantee constant-time rank. We also investigate using the same data structure to support the select queries as well. Experimental evaluation of the introduced scheme revealed that the proposed data structure consumes nearly 30%-50% less space than its alternatives by introducing less than 5% overhead, while the speed is either better or very competitive when compared with the current state-of the art implementations both in terms of rank and select.
The extensive use of online social media has highlighted the importance of privacy in the digital space. As more scientists analyse the data created in these platforms, privacy concerns have extended to data usage wit...
详细信息
Hierarchical Federated Learning (HFL) introduces intermediate aggregation layers, addressing the limitations of conventional Federated Learning (FL) in geographically dispersed environments with limited communication ...
详细信息
Recent works have shown that audio embeddings can improve automatic topic segmentation of formats such as radio shows. In this work we expand the work in that direction by showing how and which publicly available, pre...
Recent works have shown that audio embeddings can improve automatic topic segmentation of formats such as radio shows. In this work we expand the work in that direction by showing how and which publicly available, pre-trained neural audio embeddings can perform the task, without the need of any further fine-tuning of the audio encoders. The ranking of the encoders suggest that neural encoders pre-trained for speaker diarization and general purpose audio classification are the best suited to be used as features, beating non-neural baselines. We show that we can obtain perfect results on a newly created random dataset similar to the one used in previous work. We also show for the first time results on real-world data, proving that our method can be applied to actual radio shows with good results, but the choice of audio encoders is extremely important in order to achieve those. Finally, by releasing the datasets we used we make the contribution of providing the first (to our knowledge) publicly available, free of charge datasets for audio topic segmentation of media products.
暂无评论