Interpretable deep learning: interpretation, interpretability, trustworthiness, and beyond

Li, Xuhong; Xiong, Haoyi; Li, Xingjian; Wu, Xiaoyu; Zhang, Xiao; Liu, Ji; Bian, Jiang; Dou, Dejing

doi:10.1007/s10115-022-01756-8

Cited by 187 publications

(69 citation statements)

References 97 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There has been work on interpreting deep CNNs to explore their internal reasoning (X. Li et al 2021). By applying available interpretation algorithms it would be possible to identify which parts of a scene are triggering the networks' decision.…”

Section: Discussionmentioning

confidence: 99%

“…Deep neural networks are considered 'black box' models so it is difficult to say how the networks were making decisions and to what degree they were recognising different potential drivers of loss. There has been work on interpreting deep CNNs to explore their internal reasoning (Li et al, 2021). By applying interpretation algorithms, it may be possible to identify which parts of a scene are triggering the networks' decision.…”

Section: Probing the Black Boxmentioning

confidence: 99%

See 1 more Smart Citation

Using deep convolutional neural networks to forecast spatial patterns of Amazonian deforestation

et al. 2022

View full text Add to dashboard Cite

1. Tropical forests are subject to diverse deforestation pressures while their conservation is essential to achieve global climate goals. Predicting the location of deforestation is challenging due to the complexity of the natural and human systems involved but accurate and timely forecasts could enable effective planning and on-the-ground enforcement practices to curb deforestation rates. New computer vision technologies based on deep learning can be applied to the increasing volume of Earth observation data to generate novel insights and make predictions with unprecedented accuracy. 2. Here, we demonstrate the ability of deep convolutional neural networks (CNNs) to learn spatiotemporal patterns of deforestation from a limited set of freely available global data layers, including multispectral satellite imagery, the Hansen maps of annual forest change (2001-2020) and the ALOS PALSAR digital surface model, to forecast deforestation (2021). We designed four model architectures, based on 2D CNNs, 3D CNNs, and Convolutional Long Short-Term Memory (ConvLSTM) Recurrent Neural Networks (RNNs), to produce spatial maps that indicate the risk to each forested pixel (~30 m) in the landscape of becoming deforested within the next year. They were trained and tested on data from two ~80,000 km 2 tropical forest regions in the Southern Peruvian Amazon.3. The networks could predict the location of future forest loss to a high degree of accuracy (F 1 = 0.58-0.71). Our best performing model (3D CNN) had the highest pixel-wise accuracy (F 1 = 0.71) when validated on 2020 forest loss (2014-2019 training). Visual interpretation of the mapped forecasts indicated that the network could automatically discern the drivers of forest loss from the input data.For example, pixels around new access routes (e.g. roads) were assigned high risk, whereas this was not the case for recent, concentrated natural loss events (e.g. remote landslides). 4. Convolutional neural networks can harness limited time-series data to predict near-future deforestation patterns, an important step in harnessing the growing volume of satellite remote sensing data to curb global deforestation. The | 2623 Methods in Ecology and Evoluঞon BALL et al.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Probing the Black Boxmentioning

confidence: 99%

Using deep convolutional neural networks to forecast spatial patterns of Amazonian deforestation

et al. 2022

View full text Add to dashboard Cite

show abstract

“…The field of interpretable AI is itself a major research area that is crucial to gaining a better understanding of black box ML models such as DNNs. Based on the nature of the ML model, available data, and interpretation strategy, interpretable AI methods have been categorized [ 131 ]. In future work, it is imperative to determine interpretable AI methods best suited for the medical diagnosis domain.…”

Section: Discussionmentioning

confidence: 99%

Breast Cancer Dataset, Classification and Detection Using Deep Learning

Iqbal

Waqas²,

Alizadehsani

et al. 2022

Healthcare

View full text Add to dashboard Cite

Incorporating scientific research into clinical practice via clinical informatics, which includes genomics, proteomics, bioinformatics, and biostatistics, improves patients’ treatment. Computational pathology is a growing subspecialty with the potential to integrate whole slide images, multi-omics data, and health informatics. Pathology and laboratory medicine are critical to diagnosing cancer. This work will review existing computational and digital pathology methods for breast cancer diagnosis with a special focus on deep learning. The paper starts by reviewing public datasets related to breast cancer diagnosis. Additionally, existing deep learning methods for breast cancer diagnosis are reviewed. The publicly available code repositories are introduced as well. The paper is closed by highlighting challenges and future works for deep learning-based diagnosis.

show abstract

“…For interpretability, various researches focused on the mathematical level to introduce explanation to neural networks (Li et al, 2022). Such effort is important but difficult.…”

Section: Interpretabilitymentioning

confidence: 99%

Rethink Deep Learning for Clinical Natural Language Processing: Challenges and Opportunities

Luo¹

2022

Preprint

View full text Add to dashboard Cite

Deep Learning (DL) has become a popular technique in various domains to assist the human decisions, including games, news, finance, economy, biology, medicine and engineering. For clinical domain, even though statistical learning approaches have been used for clinical decision support for years, except the image related tasks, the application of deep learning mostly stays in the literature, not in the real practice. As the key component of clinical decision support system, clinical natural language processing (NLP) still relies on rule-based or statistical-based solutions while other domains have widely applied deep learning based approaches. In this paper, we provide a view on the challenges that deep learning will face for clinical NLP applications, and discuss the not yet funded opportunities from it. Overall, we want to show the broad picture about clinical NLP to deep learning researchers and encourage them to actively contribute to this field.

show abstract

Interpretable deep learning: interpretation, interpretability, trustworthiness, and beyond

Cited by 187 publications

References 97 publications

Using deep convolutional neural networks to forecast spatial patterns of Amazonian deforestation

Using deep convolutional neural networks to forecast spatial patterns of Amazonian deforestation

Breast Cancer Dataset, Classification and Detection Using Deep Learning

Rethink Deep Learning for Clinical Natural Language Processing: Challenges and Opportunities

Contact Info

Product

Resources

About