Search for: All records

Providing model explanations has gained significant popularity recently. In contrast with the traditional feature-level model explanations, concept-based explanations can provide explanations in the form of high-level human concepts. However, existing concept-based explanation methods implicitly follow a two-step procedure that involves human intervention. Specifically, they first need the human to be involved to define (or extract) the high-level concepts, and then manually compute the importance scores of these identified concepts in a post-hoc way. This laborious process requires significant human effort and resource expenditure due to manual work, which hinders their large-scale deployability. In practice, it is challenging to automatically generate the concept-based explanations without human intervention due to the subjectivity of defining the units of concept-based interpretability. In addition, due to its data-driven nature, the interpretability itself is also potentially susceptible to malicious manipulations. Hence, our goal in this paper is to free human from this tedious process, while ensuring that the generated explanations are provably robust to adversarial perturbations. We propose a novel concept-based interpretation method, which can not only automatically provide the prototype-based concept explanations but also provide certified robustness guarantees for the generated prototype-based explanations. We also conduct extensive experiments on real-world datasets to verify the desirable properties of the proposed method. 

Treatment for multiple sclerosis (MS) focuses on managing its symptoms (e.g., depression, fatigue, poor sleep quality), varying with specific symptoms experienced. Thus, for optimal treatment, there arises the need to track these symptoms. Towards this goal, there is great interest in finding their relevant phenotypes. Prior research suggests links between activities of daily living (ADLs) and MS symptoms; therefore, we hypothesize that the behavioral phenotype (revealed through ADLs) is closely related to MS symptoms. Traditional approaches to finding behavioral phenotypes which rely on human observation or controlled clinical settings are burdensome and cannot account for all genuine ADLs. Here, we present MSLife, an end-to-end, burden-free approach to digital behavioral phenotyping of MS symptoms in the wild using wearables and graph-based statistical analysis. MSLife is built upon (1) low-cost, unobtrusive wearables (i.e., smartwatches) that can track and quantify ADLs among MS patients in the wild; (2) graph-based statistical analysis that can model the relationships between quantified ADLs (i.e., digital behavioral phenotype) and MS symptoms. We design, implement, and deploy MSLife with 30 MS patients across a one-week home-based IRB-approved clinical pilot study. We use the GENEActiv smartwatch to monitor ADLs and clinical behavioral instruments to collect MS symptoms. Then we develop a graph-based statistical analysis framework to model phenotyping relationships between ADLs and MS symptoms, incorporating confounding demographic factors. We discover 102 significant phenotyping relationships (e.g., later rise times are related to increased levels of depression, history of caffeine consumption is associated with lower fatigue levels, higher relative levels of moderate physical activity are linked with decreased sleep quality). We validate their healthcare implications, using them to track MS symptoms in retrospective analysis. To our best knowledge, this is one of the first practices to digital behavioral phenotyping of MS symptoms in the wild. 

Causal inference is a critical research topic across many domains, such as statistics, computer science, education, public policy, and economics, for decades. Nowadays, estimating causal effect from observational data has become an appealing research direction owing to the large amount of available data and low budget requirement, compared with randomized controlled trials. Embraced with the rapidly developed machine learning area, various causal effect estimation methods for observational data have sprung up. In this survey, we provide a comprehensive review of causal inference methods under the potential outcome framework, one of the well-known causal inference frameworks. The methods are divided into two categories depending on whether they require all three assumptions of the potential outcome framework or not. For each category, both the traditional statistical methods and the recent machine learning enhanced methods are discussed and compared. The plausible applications of these methods are also presented, including the applications in advertising, recommendation, medicine, and so on. Moreover, the commonly used benchmark datasets as well as the open-source codes are also summarized, which facilitate researchers and practitioners to explore, evaluate and apply the causal inference methods. 

Estimating the treatment effect benefits decision making in various domains as it can provide the potential outcomes of different choices. Existing work mainly focuses on covariates with numerical values, while how to handle covariates with textual information for treatment effect estimation is still an open question. One major challenge is how to filter out the nearly instrumental variables which are the variables more predictive to the treatment than the outcome. Conditioning on those variables to estimate the treatment effect would amplify the estimation bias. To address this challenge, we propose a conditional treatment-adversarial learning based matching method (CTAM). CTAM incorporates the treatment-adversarial learning to filter out the information related to nearly instrumental variables when learning the representations, and then it performs matching among the learned representations to estimate the treatment effects. The conditional treatment-adversarial learning helps reduce the bias of treatment effect estimation, which is demonstrated by our experimental results on both semi-synthetic and real-world datasets.