Filters








115,598 Hits in 3.2 sec

Generating Token-Level Explanations for Natural Language Inference

James Thorne, Andreas Vlachos, Christos Christodoulopoulos, Arpit Mittal
2019 Proceedings of the 2019 Conference of the North  
We find that our white-box MIL-based method, while orders of magnitude faster, does not reach the same accuracy as the black-box methods.  ...  The task of Natural Language Inference (NLI) is widely modeled as supervised sentence pair classification.  ...  We compared the LIME and Anchors black-box methods against a novel, white-box Multiple Instance Learning (MIL) method and a fully supervised baseline.  ... 
doi:10.18653/v1/n19-1101 dblp:conf/naacl/ThorneVCM19 fatcat:uezggv2uxna2nemodj7um5o5y4

Generating Token-Level Explanations for Natural Language Inference [article]

James Thorne, Andreas Vlachos, Christos Christodoulopoulos, Arpit Mittal
2019 arXiv   pre-print
We find that our white-box MIL-based method, while orders of magnitude faster, does not reach the same accuracy as the black-box methods.  ...  The task of Natural Language Inference (NLI) is widely modeled as supervised sentence pair classification.  ...  We compared the LIME and Anchors black-box methods against a novel, white-box Multiple Instance Learning (MIL) method and a fully supervised baseline.  ... 
arXiv:1904.10717v1 fatcat:ylj6ion7nzc2zpjz23erqorru4

Unsupervised Black-Box Model Domain Adaptation for Brain Tumor Segmentation

Xiaofeng Liu, Chaehwa Yoo, Fangxu Xing, C.-C. Jay Kuo, Georges El Fakhri, Je-Won Kang, Jonghye Woo
2022 Frontiers in Neuroscience  
To address this issue, we propose a practical framework for UDA with a black-box segmentation model trained in the source domain only, without relying on source data or a white-box source model in which  ...  The supervised source domain training is independent of the adaptation stage in our black-box UDA setting.  ...  The unsupervised learning is combined collaboratively with the black-box source model supervision to update the target model.  ... 
doi:10.3389/fnins.2022.837646 pmid:35720708 pmcid:PMC9201342 fatcat:srpmcqkmybhedcqxu64ef7bm7u

An interpretable semi-supervised classifier using two different strategies for amended self-labeling [article]

Isel Grau, Dipankar Sengupta, Maria M. Garcia Lorenzo, Ann Nowe
2020 arXiv   pre-print
Regrettably, most successful semi-supervised classifiers do not allow explaining their outcome, thus behaving like black boxes.  ...  In this paper, we report on an extended experimental study presenting an interpretable self-labeling grey-box classifier that uses a black box to estimate the missing class labels and a white box to explain  ...  the true data distribution and improve the semi-supervised performance.  ... 
arXiv:2001.09502v2 fatcat:y2a6j54pxrhgbo5u432nksnsea

Divide to Adapt: Mitigating Confirmation Bias for Domain Adaptation of Black-Box Predictors [article]

Jianfei Yang, Xiangyu Peng, Kai Wang, Zheng Zhu, Jiashi Feng, Lihua Xie, Yang You
2022 arXiv   pre-print
Domain Adaptation of Black-box Predictors (DABP) aims to learn a model on an unlabeled target domain supervised by a black-box predictor trained on a source domain.  ...  Existing DABP approaches mostly rely on model distillation from the black-box predictor, i.e., training the model with its noisy target-domain predictions, which however inevitably introduces the confirmation  ...  Conclusion In this work, we propose BETA which learns the target domain as a partially-noisy semi-supervised learning task for black-box model adaptation.  ... 
arXiv:2205.14467v1 fatcat:kj6er6xawrdnbactrnpnq63kay

Obtaining Faithful Interpretations from Compositional Neural Networks [article]

Sanjay Subramanian, Ben Bogin, Nitish Gupta, Tomer Wolfson, Sameer Singh, Jonathan Berant, Matt Gardner
2020 arXiv   pre-print
To remedy that, we train the model with auxiliary supervision and propose particular choices for module architecture that yield much better faithfulness, at a minimal cost to accuracy.  ...  Then, the filter module takes p as input as well as the word 'black', and is meant to output high probabilities for bounding boxes with 'black dogs'.  ...  dogs in the image is equal to the number of black dogs  ... 
arXiv:2005.00724v2 fatcat:xtfi2ikjfncabon7v7zumwig44

Am I Building a White Box Agent or Interpreting a Black Box Agent? [article]

Tom Bewley
2020 arXiv   pre-print
I then discuss two independent research directions - building white box agents and interpreting black box agents - which are both coherent and worthy of attention, but must not be conflated by researchers  ...  The rule extraction literature contains the notion of a fidelity-accuracy dilemma: when building an interpretable model of a black box function, optimising for fidelity is likely to reduce performance  ...  The most obvious choice is to use samples from the state distribution encountered by the black box, d B , which gives us one measure of fidelity F B that is closest to a conventional supervised learning  ... 
arXiv:2007.01187v3 fatcat:gpff4pfr2fbunmtopil2w7zyhm

Defense Through Diverse Directions [article]

Christopher M. Bender, Yang Li, Yifeng Shi, Michael K. Reiter, Junier B. Oliva
2020 arXiv   pre-print
Unlike previous efforts in this direction, we do not rely solely on the stochasticity of network weights by minimizing the divergence between the learned parameter distribution and a prior.  ...  We demonstrate that by encouraging the network to distribute evenly across inputs, the network becomes less susceptible to localized, brittle features which imparts a natural robustness to targeted perturbations  ...  Typical methods rely on comparing white-box and black-box attacks.  ... 
arXiv:2003.10602v1 fatcat:3j6h66e4m5hpbej5guif2onmxe

An Interpretable Semi-supervised Classifier using Rough Sets for Amended Self-labeling

Isel Grau, Dipankar Sengupta, Maria M. Garcia Lorenzo, Ann Nowe
2020 2020 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE)  
However, most successful semi-supervised classifiers involve complex ensemble structures and iterative algorithms which make it difficult to explain the outcome, thus behaving like black boxes.  ...  In this paper, we build upon an interpretable self-labeling grey-box classifier that uses a black box to estimate the missing class labels and a white box to make the final predictions.  ...  Once the supervised learning is completed, the black-box component has learned a function f : L → Y , where f ∈ F , being F the hypothesis space that associates each instance with a class label.  ... 
doi:10.1109/fuzz48607.2020.9177549 dblp:conf/fuzzIEEE/GrauSLN20 fatcat:hy7sqc3ibrc7pdnr3rt5airitu

Distributed Attention for Grounded Image Captioning [article]

Nenglun Chen, Xingjia Pan, Runnan Chen, Lei Yang, Zhiwen Lin, Yuqiang Ren, Haolei Yuan, Xiaowei Guo, Feiyue Huang, Wenping Wang
2021 arXiv   pre-print
We study the problem of weakly supervised grounded image captioning.  ...  Specifically, we design a distributed attention mechanism to enforce the network to aggregate information from multiple spatially different regions with consistent semantics while generating the words.  ...  The supervised methods [12, 29, 35, 51] took bounding boxes as supervision to enforce the alignment between image regions and noun phrases, and have achieved remarkable success.  ... 
arXiv:2108.01056v1 fatcat:bflaowpm2ng5dhd2k5ms2kfm7e

Review of the Performance of Low-Cost Sensors for Air Quality Monitoring

Federico Karagulian, Maurizio Barbiere, Alexander Kotsev, Laurent Spinelle, Michel Gerboles, Friedrich Lagler, Nathalie Redon, Sabine Crunaire, Annette Borowiak
2019 Atmosphere  
A3A4 Distribution of R 2 from the comparison of SSys minute data against reference measurements. Numbers in blue and black indicate the number of open source and black box records, respectively.  ...  A4A5 Distribution of R 2 from the comparison of SSys hourly data against reference measurements. Numbers in blue and black indicate the number of open source and black box records, respectively.  ... 
doi:10.3390/atmos10090506 fatcat:nbbgj76hlbafnbwih3vlgjm764

APRICOT: A Dataset of Physical Adversarial Attacks on Object Detection [article]

A. Braunegg, Amartya Chakraborty, Michael Krumdick, Nicole Lape, Sara Leary, Keith Manville, Elizabeth Merkhofer, Laura Strickhart, Matthew Walmer
2020 arXiv   pre-print
Our analysis suggests that maintaining adversarial robustness in uncontrolled settings is highly challenging, but it is still possible to produce targeted detections under white-box and sometimes black-box  ...  We establish baselines for defending against adversarial patches through several methods, including a detector supervised with synthetic data and unsupervised methods such as kernel density estimation,  ...  and black-box attacks, and (3) several supervised and unsupervised strategies for flagging adversarial patches in real-world scenes.  ... 
arXiv:1912.08166v2 fatcat:b3uekiyoinf6zl2skwgbp2sgg4

Page 233 of POWER Vol. 105, Issue 6 [page]

1961 POWER  
JUNE 1961 PROFESSIONAL SERVICES BLACK & VEATCH Consulting Engineers Electricity—Water—Sewage—Gas—Industry Reports, Design, Supervision of Construction Investigations, Valuation and Rates 1500 Meadow Lake  ...  Consulting & Design Engineers Purchasing Construction Management TIPPETT & GEE CONSULTING ENGINEERS Mechanical—Electrical—Thermodynamics—Structural esign—Studies—Supervision Power Stations—Transmission—Distribution  ... 

NCIS: Neural Contextual Iterative Smoothing for Purifying Adversarial Perturbations [article]

Sungmin Cha, Naeun Ko, Youngjoon Yoo, Taesup Moon
2021 arXiv   pre-print
We propose a novel and effective purification based adversarial defense method against pre-processor blind white- and black-box attacks.  ...  We generate adversarial examples by ensemble transfer-based black-box attack, which can induce complete misclassification of APIs, and demonstrate that our method can be used to increase adversarial robustness  ...  Black-box attack We evaluate baselines and SSICS on two types of black-box attack.  ... 
arXiv:2106.11644v2 fatcat:cwpcbqxrs5hztbddnoszwrcwg4

Page 289 of Catholic School Journal Vol. 11, Issue 7 [page]

1911 Catholic School Journal  
Black 1 Prang No. 7 Brush Black enameled box, brush and 4 cakes of colors.  ...  Beauti- black. 15c each 20c each Price, per box................ Price, per dozen boxes 34Barclay St.  ... 
« Previous Showing results 1 — 15 out of 115,598 results