| 1. |
- Adams, Henry, et al.
(author)
-
javaPlex : A Research Software Package for Persistent (Co) Homology
- 2014
-
Conference paper (peer-reviewed)abstract
- The computation of persistent homology has proven a fundamental component of the nascent field of topological data analysis and computational topology. We describe a new software package for topological computation, with design focus on needs of the research community. This tool, replacing previous jPlex and Plex, enables researchers to access state of the art algorithms for persistent homology, cohomology, hom complexes, filtered simplicial complexes, filtered cell complexes, witness complex constructions, and many more essential components of computational topology. We describe, herewithin, the design goals we have chosen, as well as the resulting software package, and some of its more novel capabilities.
|
|
| 2. |
- Alistarh, Dan, et al.
(author)
-
The Convergence of Sparsified Gradient Methods
- 2018
-
In: Advances in Neural Information Processing Systems 31 (NIPS 2018). - : Neural Information Processing Systems (NIPS).
-
Conference paper (peer-reviewed)abstract
- Stochastic Gradient Descent (SGD) has become the standard tool for distributed training of massive machine learning models, in particular deep neural networks. Several families of communication-reduction methods, such as quantization, large-batch methods, and gradient sparsification, have been proposed to reduce the overheads of distribution. To date, gradient sparsification methods-where each node sorts gradients by magnitude, and only communicates a subset of the components, accumulating the rest locally-are known to yield some of the largest practical gains. Such methods can reduce the amount of communication per step by up to three orders of magnitude, while preserving model accuracy. Yet, this family of methods currently has no theoretical justification. This is the question we address in this paper. We prove that, under analytic assumptions, sparsifying gradients by magnitude with local error correction provides convergence guarantees, for both convex and non-convex smooth objectives, for data-parallel SGD. The main insight is that sparsification methods implicitly maintain bounds on the maximum impact of stale updates, thanks to selection by magnitude. Our analysis also reveals that these methods do require analytical conditions to converge well, justifying and complementing existing heuristics.
|
|
| 3. |
- Almstrom, Peter, et al.
(author)
-
Networked state estimation over a Gilbert-Elliot type channel
- 2009
-
In: Proceedings of the IEEE Conference on Decision and Control. ; , s. 2711-2716
-
Conference paper (peer-reviewed)abstract
- We characterize the stability and achievable performance of networked estimation under correlated packet losses described by the Gilbert-Elliot model. For scalar continuous-time linear systems, we derive closed-form expressions for the mean-square distortion of the optimal estimator. The conditions for stable mean square estimation error are equivalent to those obtained previously for stability of 'peak distortions' [3]. We study how the estimator performance depends on loss probability and loss burst length, and show that the mean-square distortion remains bounded if the average burst length does not exceeda calculated bound. The main new finding is that, once we fix the mean length of loss bursts, the average packet loss rate influences the estimator's performance but not its stability.
|
|
| 4. |
- Aminian, Behdad, et al.
(author)
-
GISOO : A virtual testbed for wireless cyber-physical systems
- 2013
-
In: Industrial Electronics Society, IECON 2013 - 39th Annual Conference of the IEEE. - : IEEE. - 9781479902248 ; , s. 5588-5593
-
Conference paper (peer-reviewed)abstract
- The increasing demand for wireless cyber-physical systems requires correct design, implementation and validation of computation, communication and control methods. Traditional simulation tools, which focus on either computation, communication or control, are insufficient when the three aspects interact. Efforts to extend the traditional tools to cover multiple domains, e.g., from simulating only control aspects to simulating both control and communication, often rely on simplistic models of a small subset of possible communication solutions. We introduce GISOO, a virtual testbed for simulation of wireless cyber-physical systems that integrates two state-of-the art simulators, Simulink and COOJA. GISOO enables users to evaluate actual embedded code for the wireless nodes in realistic cyber-physical experiments, observing the effects of both the control and communication components. In this way, a wide range of communication solutions can be evaluated without developing abstract models of their control-relevant aspects, and changes made to the networking code in simulations is guaranteed to be translated into production code without errors. A double-tank laboratory experimental setup controlled over a multi-hop relay wireless network is used to validate GISOO and demonstrate its features.
|
|
| 5. |
- Andersson, Malin, et al.
(author)
-
A Continuous-time LPV model for battery state-of-health estimation using real vehicle data
- 2020
-
In: CCTA 2020 - 4th IEEE Conference on Control Technology and Applications. - : Institute of Electrical and Electronics Engineers Inc.. ; , s. 692-698
-
Conference paper (peer-reviewed)abstract
- One approach for State-of-health estimation onboard electric vehicles is to train a data-driven virtual battery on operational data and use this model, rather than the actual battery, for performance tests. A temperature-dependent continuous-time output-error (OE) model is proposed as virtual battery and identified and validated on real operational data from electric buses. The proposed model is compared to discrete-time and parameter-invariant models and shows better performance on all data sets. In addition, the OE model structure is shown to be superior to a conventional Auto Regressive eXogenous (ARX) model for the purpose of modeling the battery voltage response. Finally, challenges regarding vehicle log data are identified and improvements to the model are suggested in order to capture observed un-modeled phenomena.
|
|
| 6. |
- António, Gonga, et al.
(author)
-
Multi-channel communication vs. adaptive routing for reliable communication in WSNs
- 2012
-
In: IPSN'12 - ACM/IEEE Proceedings of the 11th International Conference on Information Processing in Sensor Networks, (Beijing;16 - 20 April 2012). - New York, NY, USA : ACM. - 9781450312271 ; , s. 125-126
-
Conference paper (peer-reviewed)abstract
- Interference and link dynamics constitute great concerns for stability and performance of protocols in WSNs. In this paper we evaluate the impact of channel hopping and adaptive routing on delay and reliability focusing on delay critical applications.
|
|
| 7. |
- Aytekin, Arda, 1986-, et al.
(author)
-
Asynchronous incremental block-coordinate descent
- 2014
-
In: 52nd Annual Allerton Conference on Communication, Control, and Computing (Allerton). - : IEEE conference proceedings. ; , s. 19-24
-
Conference paper (peer-reviewed)abstract
- This paper studies a flexible algorithm for minimizing a sum of component functions, each of which depends on a large number of decision variables. Such formulations appear naturally in “big data” applications, where each function describes the loss estimated using the data available at a specific machine, and the number of features under consideration is huge. In our algorithm, a coordinator updates a global iterate based on delayed partial gradients of the individual objective functions with respect to blocks of coordinates. Delayed incremental gradient and delayed coordinate descent algorithms are obtained as special cases. Under the assumption of strong convexity and block coordinate-wise Lipschitz continuous partial gradients, we show that the algorithm converges linearly to a ball around the optimal value. Contrary to related proposals in the literature, our algorithm is delay-insensitive: it converges for any bounded information delay, and its step-size parameter can be chosen independently of the maximum delay bound.
|
|
| 8. |
- Aytekin, Arda, 1986-, et al.
(author)
-
Exploiting serverless runtimes for large-scale optimization
- 2019
-
In: 2019 IEEE 12th International Conference on Cloud Computing (CLOUD). - : IEEE Computer Society. ; , s. 499-501
-
Conference paper (peer-reviewed)abstract
- Serverless runtimes provide efficient and cost-effective environments for scalable computations, thanks to their event-driven and elastic nature. So far, they have mostly been used for stateless, data parallel and sporadic computations. In this work, we propose exploiting serverless runtimes to solve generic, large-scale optimization problems. To this end, we implement a parallel optimization algorithm for solving a regularized logistic regression problem, and use AWS Lambda for the compute-intensive work. We show that relative speedups up to 256 workers and efficiencies above 70% up to 64 workers can be expected.
|
|
| 9. |
- Belleschi, Marco, et al.
(author)
-
Fast power control for cross-layer optimal resource allocation in DS-CDMA wireless networks
- 2009
-
In: 2009 IEEE International Conference on Communications. ; , s. 4341-4346
-
Conference paper (peer-reviewed)abstract
- This paper presents a novel cross-layer design for joint power and end-to-end rate control optimization in DS-CDMA wireless networks, along with a detailed implementation and evaluation in the network simulator ns-2. Starting with a network utility maximization formulation of the problem, we derive distributed power control, transport rate and queue management schemes that jointly achieve the optimal network operation. Our solution has several attractive features compared to alternatives: it adheres to the natural time-scale separation between rapid power control updates and slower end-to-end rate adjustments, and uses simplified power control mechanisms with reduced signalling requirements. We argue that these features are critical for a successful real-world implementation. To validate these claims, we present a detailed implementation of a crosslayer adapted networking stack for DS-CSMA ad-hoc networks in ns-2. We describe several critical issues that arise in the implementation, but are typically neglected in the theoretical protocol design, and evaluate the alternatives in extensive simulations.
|
|
| 10. |
- Bengtsson, Stefan, 1961, et al.
(author)
-
Wafer bonding: A flexible way to manufacture SOI materials for high performance applications
- 2004
-
In: Proc. 19th Symp. on Microelectronics Technology and Devices. ; 3, s. 241-
-
Conference paper (peer-reviewed)abstract
- An overview is given on the use of wafer bonding for formation of Silicon-On-Insulator (SOI) materials for high performance applications. Recent developments in wafer bonding and available techniques for formation of thin semiconductor films is presented. Furthermore, a review is given on results in use of wafer bonding for formation of advanced SOI-materials. Finally, a more detailed discussion is given on the use of wafer bonding for manufacture of SOI-materials intended for high-frequency applications and SOI-materials with films of electrically insulating but highly thermally conductive materials as buried insulators.
|
|
