TRACER: A Framework for Facilitating Accurate and Interpretable Analytics for High Stakes Applications

Zheng, Kaiping; Cai, Shaofeng; Chua, Horng Ruey; Wang, Wei; Ngiam, Kee Yuan; Ooi, Beng Chin

doi:10.1145/3318464.3389720

Cited by 16 publications

(11 citation statements)

References 60 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We proposed a novel time-invariant and time-variant (TITV) model to facilitate more accurate and interpretable analytics in AKI prediction based on the collaboration of 3 modules [17] (Figure 1). In the time-invariant module, an abstract representation was calculated with the data in the entire feature window, denoting each feature's importance shared across time (ie, time-invariant feature importance).…”

Section: Analyticsmentioning

confidence: 99%

Health Care Analytics With Time-Invariant and Time-Variant Feature Importance to Predict Hospital-Acquired Acute Kidney Injury: Observational Longitudinal Study

et al. 2021

Self Cite

View full text Add to dashboard Cite

Background Acute kidney injury (AKI) develops in 4% of hospitalized patients and is a marker of clinical deterioration and nephrotoxicity. AKI onset is highly variable in hospitals, which makes it difficult to time biomarker assessment in all patients for preemptive care. Objective The study sought to apply machine learning techniques to electronic health records and predict hospital-acquired AKI by a 48-hour lead time, with the aim to create an AKI surveillance algorithm that is deployable in real time. Methods The data were sourced from 20,732 case admissions in 16,288 patients over 1 year in our institution. We enhanced the bidirectional recurrent neural network model with a novel time-invariant and time-variant aggregated module to capture important clinical features temporal to AKI in every patient. Time-series features included laboratory parameters that preceded a 48-hour prediction window before AKI onset; the latter’s corresponding reference was the final in-hospital serum creatinine performed in case admissions without AKI episodes. Results The cohort was of mean age 53 (SD 25) years, of whom 29%, 12%, 12%, and 53% had diabetes, ischemic heart disease, cancers, and baseline eGFR <90 mL/min/1.73 m2, respectively. There were 911 AKI episodes in 869 patients. We derived and validated an algorithm in the testing dataset with an AUROC of 0.81 (0.78-0.85) for predicting AKI. At a 15% prediction threshold, our model generated 699 AKI alerts with 2 false positives for every true AKI and predicted 26% of AKIs. A lowered 5% prediction threshold improved the recall to 60% but generated 3746 AKI alerts with 6 false positives for every true AKI. Representative interpretation results produced by our model alluded to the top-ranked features that predicted AKI that could be categorized in association with sepsis, acute coronary syndrome, nephrotoxicity, or multiorgan injury, specific to every case at risk. Conclusions We generated an accurate algorithm from electronic health records through machine learning that predicted AKI by a lead time of at least 48 hours. The prediction threshold could be adjusted during deployment to optimize recall and minimize alert fatigue, while its precision could potentially be augmented by targeted AKI biomarker assessment in the high-risk cohort identified.

show abstract

Section: Analyticsmentioning

confidence: 99%

Health Care Analytics With Time-Invariant and Time-Variant Feature Importance to Predict Hospital-Acquired Acute Kidney Injury: Observational Longitudinal Study

et al. 2021

Self Cite

View full text Add to dashboard Cite

show abstract

“…Many existing works [12,48] capture cross features in an implicit manner with DNNs. However, modeling multiplicative interaction implicitly with DNNs requires a substantial number of hidden units [2,6,25], which makes the modeling process inefficient and less interpretable [13,18,61].…”

Section: Related Workmentioning

confidence: 99%

“…Particularly, the attention mechanism [4] is widely adopted to facilitate the interpretability of deep models by visualizing the attention weights. With the attention mechanism integrated into the model design, many studies [14,36,61] manage to achieve interpretable healthcare analytics. Specifically, Dipole [36] supports the visit-level interpretation in the diagnosis prediction with three attention mechanisms.…”

Section: Related Workmentioning

confidence: 99%

“…Specifically, Dipole [36] supports the visit-level interpretation in the diagnosis prediction with three attention mechanisms. RETAIN [14] and TRACER [61] can support both the visit-level and the feature-level interpretation. However, one inherent limitation of most existing methods is that their interpretability is built upon single input features, while disregarding feature interactions that are essential for relational analytics.…”

Section: Related Workmentioning

confidence: 99%

“…In particular, the deployment of learning models in real-world applications has highlighted the need for not only accuracy, but also efficiency and interpretability [7,18,38,46]. Notably, understanding the general behavior and the whole logic of the learning model (global interpretability) and providing reasons for making a specific decision (local interpretability) [18,46] is of paramount importance for high-stakes applications in critical decisions making, e.g., healthcare or finance [61]. Despite their great predictive power, many "black box" models such as DNNs model inputs in an implicit way, which is inexplicable and sometimes can learn unintended patterns [53,60].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

ARM-Net: Adaptive Relation Modeling Network for Structured Data

Cai

Zheng

Chen

et al. 2021

Proceedings of the 2021 International Conference on Management of Data

Self Cite

View full text Add to dashboard Cite

Relational databases are the standard for storing and querying structured data, and extracting insights from structured data requires advanced analytics. Deep neural networks (DNNs) have achieved super-human prediction performance in particular data types, e.g., images. However, existing DNNs may not produce meaningful results when applied to structured data. The reason is that there are correlations and dependencies across combinations of attribute values in a table, and these do not follow simple additive patterns that can be easily mimicked by a DNN. The number of possible such "cross features" is combinatorial, making them computationally prohibitive to model. Furthermore, the deployment of learning models in real-world applications has also highlighted the need for interpretability, especially for high-stakes applications, which remains another issue of concern to DNNs.In this paper, we present ARM-Net, an adaptive relation modeling network tailored for structured data, and a lightweight framework ARMOR based on ARM-Net for relational data analytics. The key idea is to model feature interactions with cross features selectively and dynamically, by first transforming the input features into exponential space, and then determining the interaction order and interaction weights adaptively for each cross feature. We propose a novel sparse attention mechanism to dynamically generate the interaction weights given the input tuple, so that we can explicitly model cross features of arbitrary orders with noisy features filtered selectively. Then during model inference, ARM-Net can specify the cross features being used for each prediction for higher accuracy and better interpretability. Our extensive experiments on real-world datasets demonstrate that ARM-Net consistently outperforms existing models and provides more interpretable predictions for data-driven decision making.

show abstract

Explainable AI for clinical and remote health applications: a survey on tabular and time series data

Martino

Delmastro

2022

Artif Intell Rev

View full text Add to dashboard Cite

Nowadays Artificial Intelligence (AI) has become a fundamental component of healthcare applications, both clinical and remote, but the best performing AI systems are often too complex to be self-explaining. Explainable AI (XAI) techniques are defined to unveil the reasoning behind the system’s predictions and decisions, and they become even more critical when dealing with sensitive and personal health data. It is worth noting that XAI has not gathered the same attention across different research areas and data types, especially in healthcare. In particular, many clinical and remote health applications are based on tabular and time series data, respectively, and XAI is not commonly analysed on these data types, while computer vision and Natural Language Processing (NLP) are the reference applications. To provide an overview of XAI methods that are most suitable for tabular and time series data in the healthcare domain, this paper provides a review of the literature in the last 5 years, illustrating the type of generated explanations and the efforts provided to evaluate their relevance and quality. Specifically, we identify clinical validation, consistency assessment, objective and standardised quality evaluation, and human-centered quality assessment as key features to ensure effective explanations for the end users. Finally, we highlight the main research challenges in the field as well as the limitations of existing XAI methods.

show abstract

TRACER: A Framework for Facilitating Accurate and Interpretable Analytics for High Stakes Applications

Cited by 16 publications

References 60 publications

Health Care Analytics With Time-Invariant and Time-Variant Feature Importance to Predict Hospital-Acquired Acute Kidney Injury: Observational Longitudinal Study

Health Care Analytics With Time-Invariant and Time-Variant Feature Importance to Predict Hospital-Acquired Acute Kidney Injury: Observational Longitudinal Study

ARM-Net: Adaptive Relation Modeling Network for Structured Data

Explainable AI for clinical and remote health applications: a survey on tabular and time series data

Contact Info

Product

Resources

About