Filters








7,492 Hits in 4.4 sec

Abstract Compilation for Verification of Numerical Accuracy Properties [article]

Maxime Jacquemin, Fonenantsoa Maurica, Nikolai Kosmatov, Julien Signoles, Franck Védrine
2019 arXiv   pre-print
This paper describes an original framework combining different solutions for numerical accuracy.  ...  Verification of numerical accuracy properties in modern software remains an important and challenging task.  ...  The authors thank Romain Soulat and Thales Research & Technology for providing case studies and participation in the evaluation.  ... 
arXiv:1911.10930v1 fatcat:ulxdanrqvvd3lk7apseqmcjimq

SAFE-PDF: Robust Detection of JavaScript PDF Malware Using Abstract Interpretation [article]

Alexander Jordan and François Gauthier and Behnaz Hassanshahi and David Zhao
2018 arXiv   pre-print
A comparison with two state-of-the-art PDF malware detection tools shows that our conservative abstract interpretation approach achieves similar accuracy, while being more resilient to evasion attacks.  ...  In contrast, abstract interpretation is oblivious to both types of evasions.  ...  Acknowledgment The authors would like to thank Phil Boutros and Joe Keslin from the Oracle Clean Content team for their support.  ... 
arXiv:1810.12490v1 fatcat:w3y3qvejonhelekdwnepotwggi

Keep your Distance: Determining Sampling and Distance Thresholds in Machine Learning Monitoring [article]

Al-Harith Farhad, Ioannis Sorokos, Andreas Schmidt, Mohammed Naveed Akram, Koorosh Aslansefat, Daniel Schneider
2022 arXiv   pre-print
For components whose environments are open and variable, for instance those found in autonomous vehicles, it is therefore important to monitor their operational situation to determine its distance from  ...  SafeML is a model-agnostic approach for performing such monitoring, using distance measures based on statistical testing of the training and operational datasets.  ...  Farhad et al. eral Ministry for Economic Affairs and Climate Action (BMWK) within the research project "FabOS" under grant 01MK20010A.  ... 
arXiv:2207.05078v1 fatcat:bswbrbjaizdi7auyfpkosxx6b4

Fast, robust and non-convex subspace recovery

Gilad Lerman, Tyler Maunu
2017 Information and Inference A Journal of the IMA  
Numerical experiments on synthetic and real data demonstrate its competitive speed and accuracy.  ...  This work presents a fast and non-convex algorithm for robust subspace recovery.  ...  with processing and interpreting the astronomy data; Teng Zhang for useful comments on earlier versions of this manuscript and helpful discussions; and Nati Srebro for encouraging us to write up and submit  ... 
doi:10.1093/imaiai/iax012 fatcat:b4jx3ylaonconoqf5vkslmp3ji

SpecRepair: Counter-Example Guided Safety Repair of Deep Neural Networks [article]

Fabian Bauer-Marquart, David Boetius, Stefan Leue, Christian Schilling
2022 arXiv   pre-print
The results show that SpecRepair is more successful in producing safe DNNs than comparable methods, has a shorter runtime, and produces safe DNNs while preserving their classification accuracy.  ...  We evaluate SpecRepair's effectiveness on the ACAS Xu benchmark, a DNN-based controller for unmanned aircraft, and two image classification benchmarks.  ...  Acknowledgments This research was partly supported by DIREC -Digital Research Centre Denmark and the Villum Investigator Grant S4OS. Bibliography  ... 
arXiv:2106.01917v5 fatcat:x4aiko45fbgujnd7inxbeudgr4

Living with Uncertainty in the Age of Runtime Models [chapter]

Holger Giese, Nelly Bencomo, Liliana Pasquale, Andres J. Ramirez, Paola Inverardi, Sebastian Wätzoldt, Siobhán Clarke
2014 Lecture Notes in Computer Science  
A runtime model is a dynamic knowledge base that abstracts useful information about the system, its operational context and the extent to which the system meets its stakeholders' needs.  ...  To this end, we introduce a well-suited terminology about models, runtime models and uncertainty and present a state-of-the-art summary on model-based techniques for addressing uncertainty both at development-and  ...  In general, it is expected for a model to provide an acceptable degree of accuracy and precision.  ... 
doi:10.1007/978-3-319-08915-7_3 fatcat:pqohcufa3vgojexrey75lhuv6m

How Many Bits Does it Take to Quantize Your Neural Network? [chapter]

Mirco Giacobbe, Thomas A. Henzinger, Mathias Lechner
2020 Lecture Notes in Computer Science  
We show that neither robustness nor nonrobustness are monotonic with changing the number of bits for the representation and, also, neither are preserved by quantization from a realnumbered network.  ...  We demonstrate that, compared to our method, existing methods for the analysis of real-numbered networks often derive false conclusions about their quantizations, both when determining robustness and when  ...  for Business, Energy & Industrial Strategy (BEIS), and Innovate UK under the HICLASS project (113213).  ... 
doi:10.1007/978-3-030-45237-7_5 fatcat:pcfd7jgqs5bendlhzjcan3qucm

On Using Retrained and Incremental Machine Learning for Modeling Performance of Adaptable Software: An Empirical Comparison [article]

Tao Chen
2019 arXiv   pre-print
has gained momentum for evaluating, understanding and predicting software performance, which facilitates better informed self-adaptations.  ...  the existing model and tunes it using one newly arrival data sample.  ...  This can ensure good numeric stability, which in turn, significantly improves the prediction accuracy [31] [10] .  ... 
arXiv:1903.10614v1 fatcat:ewcfvzggarct3p4ilfcof3kw3m

Robust Estimation of Neural Signals in Calcium Imaging

Hakan Inan, Murat A. Erdogdu, Mark J. Schnitzer
2017 Neural Information Processing Systems  
Using the theory of M-estimation, we derive a minimax optimal robust loss, and also find a simple and practical optimization routine for this loss with provably fast convergence.  ...  In this work, we proceed in a new direction and propose to extract cells and their activity using robust statistical estimation.  ...  Acknowledgements We gratefully acknowledge support from DARPA and technical assistance from Biafra Ahanonu, Lacey Kitch, Yaniv Ziv, Elizabeth Otto and Margaret Carr.  ... 
dblp:conf/nips/InanES17 fatcat:rofzboncnna5bm2eqvffnauyjy

Answer Set Programming Modulo `Space-Time' [article]

Carl Schultz, Mehul Bhatt, Jakob Suchan, Przemysław Wałęga
2018 arXiv   pre-print
We present an empirical evaluation (with scalability and robustness results), and include diverse application examples involving interpretation and control tasks.  ...  Supported are capabilities for mixed qualitative-quantitative reasoning, consistency checking, and inferring compositions of space-time relations; these capabilities combine and synergise for applications  ...  Figure 2 : 2 Qualitative Abstractions for Regions of Space.  ... 
arXiv:1805.06861v1 fatcat:j5eulz4egnbqlahckeq5wrl3di

A comparative evaluation of systems for scalable linear algebra-based analytics

Anthony Thomas, Arun Kumar
2018 Proceedings of the VLDB Endowment  
All of our code and data scripts are available for download at https://adalabucsd.github.io/slab.html.  ...  But implementing new scalable algorithms in low-level languages is a painful process, especially for enterprise and scientific users.  ...  We thank the developers of MADlib, SciDB, and SystemML for helpful conversations about their respective systems.  ... 
doi:10.14778/3275366.3284963 fatcat:6knwmwerlzdy5a7fxqygg3knvi

Verifying Robustness of Gradient Boosted Models [article]

Gil Einziger, Maayan Goldstein, Yaniv Sa'ar, Itai Segall
2019 arXiv   pre-print
We extensively evaluate VeriGB on publicly available datasets and demonstrate a capability for verifying large models.  ...  VeriGB encodes the model and the robustness property as an SMT formula, which enables state of the art verification tools to prove the model's robustness.  ...  Such success gave rise to diverse verification methods such as model checking, termination analysis, and abstract interpretation.  ... 
arXiv:1906.10991v1 fatcat:hcok4igcprgjrcqi4iqtfkdkny

Verifying Robustness of Gradient Boosted Models

Gil Einziger, Maayan Goldstein, Yaniv Sa'ar, Itai Segall
2019 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
We extensively evaluate VERIGB on publicly available datasets and demonstrate a capability for verifying large models.  ...  introduces VERIGB, a tool for quantifying the robustness of gradient boosted models.  ...  Such success gave rise to diverse verification methods such as model checking, termination analysis, and abstract interpretation.  ... 
doi:10.1609/aaai.v33i01.33012446 fatcat:dfvjf3ylc5eu7ilvxfap32cc6u

The Tau Parallel Performance System

Sameer S. Shende, Allen D. Malony
2006 The international journal of high performance computing applications  
Flexibility and portability in empirical methods and processes are influenced primarily by the strategies available for instrumentation and measurement, and how effectively they are integrated and composed  ...  This paper presents the TAU (Tuning and Analysis Utilities) parallel performance system and describe how it addresses diverse requirements for performance observation and analysis.  ...  The TAU project has benefited from the contributions of many project staff and graduate students. We would like to recognize in particular those of Robert Bell  ... 
doi:10.1177/1094342006064482 fatcat:tu5rcme47bctdgbrsdq2hzahp4

Combinatorial Testing for Deep Learning Systems [article]

Lei Ma, Fuyuan Zhang, Minhui Xue, Bo Li, Yang Liu, Jianjun Zhao, Yadong Wang
2018 arXiv   pre-print
The main challenge of testing such systems is that its runtime state space is too large: if we view each neuron as a runtime state for DL, then a DL system often contains massive states, rendering testing  ...  Adopting testing techniques could help to evaluate the robustness of a DL system and therefore detect vulnerabilities at an early stage.  ...  Acknowledgement This research is partially supported by a recently awarded grant Robust Deep Learning and its Application to high Confidence Medical Diagnosis, which is a pivotal sub-project of Chinese  ... 
arXiv:1806.07723v1 fatcat:jbnswfrkh5fjjdgrbiwmloyphi
« Previous Showing results 1 — 15 out of 7,492 results