SLR(k) covering for LR(k) grammars

Lee, Myung-Joon; Choe, Kwang-Moo

doi:10.1016/0020-0190(91)90073-q

Cited by 7 publications

(16 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…One way to directly assess epistemic uncertainty is to calculate the distance between training and testing activations. As a model is unlikely to produce reasonable outputs for features far from any seen during training, this is a reliable signal for bad model performance ( Lee et al, 2018b ).…”

Section: Methodsmentioning

confidence: 99%

“…Model activations have covariance, and they do not necessarily resemble the mode for high-dimensional spaces ( Wei et al, 2015 ), so the Euclidean distance is not appropriate for identifying unusual activation patterns. Instead, inspired by the work of Lee et al (2018b) , we make use of the Mahalanobis distance

, which rescales samples into a space without covariance. Fig.…”

Section: Methodsmentioning

confidence: 99%

“…We project these to a lower-dimensional feature space and estimate a multi-variate Gaussian with the training data. During inference, we detect samples with a high Mahalanobis distance to this distribution, which is suitable for quantifying differences in the latent space ( Lee et al, 2018b , Çallı et al, 2019 ).…”

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

Distance-based detection of out-of-distribution silent failures for Covid-19 lung lesion segmentation

González

Gotkowski

Fuchs

et al. 2022

Medical Image Analysis

View full text Add to dashboard Cite

Section: Methodsmentioning

confidence: 99%

, which rescales samples into a space without covariance. Fig.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Distance-based detection of out-of-distribution silent failures for Covid-19 lung lesion segmentation

González

Gotkowski

Fuchs

et al. 2022

Medical Image Analysis

View full text Add to dashboard Cite

“…Softmax-based classifiers are known to generate overconfident posterior distributions when presented with out-of-distribution detection (ODD) data. To address this issue, a generative classifier, specifically Gaussian discriminant analysis, and a Mahalanobis distance-based framework were utilized to obtain confidence scores, as described by Lee et al (45). To evaluate the performance of this approach, 164 KSα sequences were designated ODD data, and 164 KSβ sequences were designated in-distribution (ID) data.…”

Section: Detection Of T2pks With Potentially New Skeletonsmentioning

confidence: 99%

“…However, such models may demonstrate overconfidence in discerning novel KSβ sequences in the open world(27). To address this concern, an ODD framework based on the Mahalanobis distance was implemented for multiclass novelty detection(45). Notably, certain samples (from 2597 KSβ sequences) are proximal to the labeled data (from 164 KSβ sequences) because such labeled T2PKs with entirely novel carbon skeletons have only been discovered in recent years, such as formicamycin(36) and dendrubin(46).…”

mentioning

confidence: 99%

Protein language model-based end-to-end type II polyketide prediction without sequence alignment

Qin

Zhang

Huang

et al. 2023

Preprint

View full text Add to dashboard Cite

Natural products are important sources for drug development, and the precise prediction of their structures assembled by modular proteins is an area of great interest. In this study, we introduce DeepT2, an end-to-end, cost-effective, and accurate machine learning platform to accelerate the identification of type II polyketides (T2PKs), which represent a significant portion of the natural product world. Our algorithm is based on advanced natural language processing models and utilizes the core biosynthetic enzyme, chain length factor (CLF or KSbeta), as computing inputs. The process involves sequence embedding, data labeling, classifier development, and novelty detection, which enable precise classification and prediction directly from KSbeta without sequence alignments. Combined with metagenomics and metabolomics, we evaluated the ability of DeepT2 and found this model could easily detect and classify KSbeta either as a single sequence or a mixture of bacterial genomes, and subsequently identify the corresponding T2PKs in a labeled categorized class or as novel. Our work highlights deep learning as a promising framework for genome mining and therefore provides a meaningful platform for discovering medically important natural products.

show abstract

Enhancing Open-World Bacterial Raman Spectra Identification by Feature Regularization for Improved Resilience against Unknown Classes

Balytskyi,

Kalashnyk,

Hubenko

et al. 2023

Preprint

View full text Add to dashboard Cite

The combination of Deep Learning techniques and Raman spectroscopy shows great potential offering precise and prompt identification of pathogenic bacteria in clinical settings. However, the traditional closed-set classification approaches assume that all test samples belong to one of the known pathogens, and their applicability is limited since the clinical environment is inherently unpredictable and dynamic, unknown or emerging pathogens may not be included in the available catalogs. We demonstrate that the current state-of-the-art Neural Networks identifying pathogens through Raman spectra are vulnerable to unknown inputs, resulting in an uncontrollable false positive rate. To address this issue, first, we developed a novel ensemble of ResNet architectures combined with the attention mechanism which outperforms existing closed-world methods, achieving an accuracy of 87.8 ± 0.1% compared to the best available model's accuracy of 86.7 ± 0.4%. Second, through the integration of feature regularization by the Objectosphere loss function, our model achieves both high accuracy in identifying known pathogens from the catalog and effectively separates unknown samples drastically reducing the false positive rate. Finally, the proposed feature regularization method during training significantly enhances the performance of out-of-distribution detectors during the inference phase improving the reliability of the detection of unknown classes. Our novel algorithm for Raman spectroscopy enables the detection of unknown, uncatalogued, and emerging pathogens providing the flexibility to adapt to future pathogens that may emerge, and has the potential to improve the reliability of Raman-based solutions in dynamic operating environments where accuracy is critical, such as public safety applications.

show abstract

SLR(k) covering for LR(k) grammars

Cited by 7 publications

References 0 publications

Distance-based detection of out-of-distribution silent failures for Covid-19 lung lesion segmentation

Distance-based detection of out-of-distribution silent failures for Covid-19 lung lesion segmentation

Protein language model-based end-to-end type II polyketide prediction without sequence alignment

Enhancing Open-World Bacterial Raman Spectra Identification by Feature Regularization for Improved Resilience against Unknown Classes

Contact Info

Product

Resources

About