Filters








971 Hits in 3.8 sec

Synthesizing Pareto-Optimal Interpretations for Black-Box Models [article]

Hazem Torfah, Shetal Shah, Supratik Chakraborty, S. Akshay, Sanjit A. Seshia
2021 arXiv   pre-print
We present a new multi-objective optimization approach for synthesizing interpretations that "explain" the behavior of black-box machine learning models.  ...  For a given black-box, our approach yields a set of Pareto-optimal interpretations with respect to the correctness and explainability measures.  ...  We would also like to express our gratitude to the anonymous reviewers for their in-depth reviews, constructive suggestions and various pointers.  ... 
arXiv:2108.07307v1 fatcat:nuoghuyb6jeihfndntqd6sb7mm

Synthesizing Pareto-Optimal Interpretations for Black-Box Models [article]

Hazem Torfah, Shetal Shah, Supratik Chakraborty, S. Akshay, Sanjit A. Seshia
2021
We present a new multi-objective optimization approach for synthesizing interpretations that "explain" the behavior of black-box machine learning models.  ...  For a given black-box, our approach yields a set of Pareto-optimal interpretations with respect to the correctness and explainability measures.  ...  We would also like to express our gratitude to the anonymous reviewers for their in-depth reviews, constructive suggestions and various pointers.  ... 
doi:10.34727/2021/isbn.978-3-85448-046-4_24 fatcat:l2lp6a7xzfgq7jytaxlt45nsbu

Practical Design Space Exploration [article]

Luigi Nardi and David Koeplinger and Kunle Olukotun
2019 arXiv   pre-print
The proposed methodology follows a white-box model which is simple to understand and interpret (unlike, for example, neural networks) and can be used by the user to better understand the results of the  ...  Our results show that HyperMapper 2.0 provides better Pareto fronts compared to state-of-the-art baselines, with better or competitive hypervolume indicator and with 8x improvement in sampling budget for  ...  This problem is referred to in the literature as derivative-free optimization (DFO) [10, 24] , also known as black-box optimization [12] and, in the computer systems community, as design-space exploration  ... 
arXiv:1810.05236v3 fatcat:65la5kczxrbtvchokqp6nqivuq

Adversarial Black-Box Attacks on Automatic Speech Recognition Systems Using Multi-Objective Evolutionary Optimization

Shreya Khare, Rahul Aralikatte, Senthil Mani
2019 Interspeech 2019  
Both black-box and white-box approaches have been used to either replicate the model itself or to craft examples which cause the model to fail.  ...  In this work, we propose a framework which uses multi-objective evolutionary optimization to perform both targeted and un-targeted blackbox attacks on Automatic Speech Recognition (ASR) systems.  ...  On the other hand, Black Box attacks, where the attacker only has access to the inputoutput pairs of the model, are more probable.  ... 
doi:10.21437/interspeech.2019-2420 dblp:conf/interspeech/KhareAM19 fatcat:xag6wuxmn5a35bpugadyg6mkjq

Towards automated design of corrosion resistant alloy coatings with an autonomous scanning droplet cell [article]

Brian DeCost, Howie Joress, Suchismita Sarker, Apurva Mehta, Jason Hattrick-Simpers
2022 arXiv   pre-print
This motivates a close coupling between autonomous research platforms and scientific machine learning methodology that blends mechanistic physical models and black box machine learning models.  ...  We present an autonomous scanning droplet cell platform designed for on-demand alloy electrodeposition and real-time electrochemical characterization for investigating the corrosion-resistance properties  ...  Code required for reproducing the analyses in the manuscript will be made available in a separate repository.  ... 
arXiv:2203.17049v1 fatcat:jp5td2shcbdwrdpdep4ia5lzqa

MBSE for Satellite Communication System Architecting

Su Gao, Wei Cao, Luhai Fan, Jihong Liu
2019 IEEE Access  
System black-box analysis and white-box logical decomposition are further realized.  ...  The logical architecture is then partitioned for physical implementation and system optimization is carried out to give architecting suggestions.  ...  For top-level models (mission scenario and black-box model), requirements, derived mission scenarios and use cases are considered.  ... 
doi:10.1109/access.2019.2952889 fatcat:hyq665gudbd2nggcr57dnxglzy

CAD Tool Design Space Exploration via Bayesian Optimization [article]

Yuzhe Ma, Ziyang Yu, Bei Yu
2019 arXiv   pre-print
It is based on Bayesian optimization which is a promising technique for optimizing black-box functions that are expensive to evaluate.  ...  Gaussian process regression is leveraged as the surrogate model in Bayesian optimization framework. In this work, we use 64-bit prefix adder design as a case study.  ...  The regression model will be guided to explore different best solutions which are Pareto-optimal.  ... 
arXiv:1912.06460v1 fatcat:xjyxzxg3jrecdpk4eoanjvbfzu

A Survey of Recent Trends in Multiobjective Optimal Control—Surrogate Models, Feedback Control and Objective Reduction

Sebastian Peitz, Michael Dellnitz
2018 Mathematical and Computational Applications  
Besides classical meta-modeling techniques for multiobjective optimization, a promising alternative for control problems is to introduce a surrogate model for the system dynamics.  ...  The task in multiobjective optimization and multiobjective optimal control is therefore to compute the set of optimal compromises (the Pareto set) between the conflicting objectives.  ...  for solving MOPs due to the applicability to very complex problems and being easy to use in a black box fashion.  ... 
doi:10.3390/mca23020030 fatcat:zixdhh4uhvbufcxbbnmyf4uxji

Data mining applied for earthworks optimisation of a toll road construction project

Andri Irfan Rifai, Yusuf Latief, Leni Sagita Rianti, P. Hajek, A.L. Han, S. Kristiawan, W.T. Chan, M.b. Ismail, B.S. Gan, R. Sriravindrarajah, B.A. Hidayat
2018 MATEC Web of Conferences  
Data on earthworks can be utilised as a knowledge base and processed via a data mining approach, the results of which form the basis for interpretation and productivity predictions.  ...  This work aims to present a practical alternative for the optimisation of earthworks.  ...  We would also like to thank people for working at Universitas Indonesia.  ... 
doi:10.1051/matecconf/201819504019 fatcat:xtsq2ird5famllovq4gx24df4a

Multi-objective Bayesian optimization of ferroelectric materials with interfacial control for memory and energy storage applications [article]

Arpan Biswas, Anna N. Morozovska, Maxim Ziatdinov, Eugene A. Eliseev, Sergei V. Kalinin
2021 arXiv   pre-print
MOBO is a low computational cost optimization tool for expensive multi-objective functions, where we update posterior surrogate Gaussian process models from prior evaluations, and then select future evaluations  ...  to required Pareto front.  ...  It has been considered as a low computationally cost global optimization tool for design problems having expensive black-box objective functions.  ... 
arXiv:2108.12889v1 fatcat:5c6icdkfb5gkjejtoewsijbqpq

Plausible Counterfactuals: Auditing Deep Learning Classifiers with Realistic Adversarial Examples [article]

Alejandro Barredo-Arrieta, Javier Del Ser
2020 arXiv   pre-print
Unfortunately, the black-box nature of Deep Learning models has posed unanswered questions about what they learn from data.  ...  This is indeed the angular stone for this research work: to undertake an adversarial analysis of a Deep Learning model.  ...  ACKNOWLEDGMENTS The authors would like to thank the Basque Government for its support through the EMAITEK and ELKARTEK funding programs.  ... 
arXiv:2003.11323v1 fatcat:77zqvcd6zfgu7jpngubo6gjppq

A methodology for co-simulation-based optimization of biofabrication protocols [article]

Leonardo Giannantoni, Roberta Bardini, Stefano Di Carlo
2022 bioRxiv   pre-print
This paper proposes a novel co-simulation-based optimization methodology for the systematic design of protocols for cell culture and biofabrication.  ...  Trial-and-error methods are costly and yield only incremental innovation, starting from sub-optimal and poorly represented existing processes.  ...  Also, in this case, white-box is preferable to black-box modeling since interpretability and explainability of OvS results build their relevance for the design of an actual biofabrication process.  ... 
doi:10.1101/2022.01.28.478198 fatcat:lspbtaq5djftbacxjxrkbfdaue

Task concurrency management methodology to schedule the MPEG4 IM1 player on a highly parallel processor platform

Chun Wong, Paul Marchal, Peng Yang
2001 Proceedings of the ninth international symposium on Hardware/software codesign - CODES '01  
Starting from the original "standard" specification, we extract the concurrency originally hidden by implementation decisions in a "grey-box" model.  ...  These curves will be used to get globally optimized design decisions when combining subsystems into one complete system or to be used by a dynamic scheduler.  ...  For Combination 2 and 3, we have derived a Pareto-optimal energy vs time-budget curve for each of them.  ... 
doi:10.1145/371636.371712 dblp:conf/codes/WongMY01 fatcat:bqqwstwvljg6rjgohe53weu2aq

Infinity Sensor: Temperature Sensing in GaN Power Devices using Peak di/dt

Jianjing Wang, Mohammad H. Hedayati, Dawei Liu, Salah-Eddine Adami, Harry C. P. Dymond, Jeremy J. O. Dalton, Bernard H. Stark
2018 2018 IEEE Energy Conversion Congress and Exposition (ECCE)  
and retrofits, possible methodology development for evaluation and synthesis, and the importance of good modeling practice.  ...  The theoretical basis of the most commonly-used multi-objective techniques and recent developments are given to offer high-quality Pareto front for decision makers, with an emphasis on evolutionary algorithms  ...  constraint in the associated black-box, as shown in Figure 8 .  ... 
doi:10.1109/ecce.2018.8558287 fatcat:6dtqnvu3tjef5nih4jqtztot34

A Review of Evaluation, Optimization and Synthesis of Energy Systems: Methodology and Application to Thermal Power Plants

Ligang Wang, Zhiping Yang, Shivom Sharma, Alberto Mian, Tzu-En Lin, George Tsatsaronis, François Maréchal, Yongping Yang
2018 Energies  
and retrofits, possible methodology development for evaluation and synthesis, and the importance of good modeling practice.  ...  The theoretical basis of the most commonly-used multi-objective techniques and recent developments are given to offer high-quality Pareto front for decision makers, with an emphasis on evolutionary algorithms  ...  constraint in the associated black-box, as shown in Figure 8 .  ... 
doi:10.3390/en12010073 fatcat:br25krgnzrb4jcvuzcjkiruq6e
« Previous Showing results 1 — 15 out of 971 results