A5026901076- Advanced Causal Inference Techniques
- Statistical Methods and Inference
- Statistical Methods and Bayesian Inference
- IoT and Edge/Fog Computing
- Statistical Methods in Clinical Trials
- Age of Information Optimization
- Health Systems, Economic Evaluations, Quality of Life
- Cloud Computing and Resource Management
- Machine Learning and ELM
- Natural Language Processing Techniques
- Topic Modeling
- Radiation Detection and Scintillator Technologies
- Privacy-Preserving Technologies in Data
- Advanced Data Processing Techniques
- Advanced Statistical Process Monitoring
- Rough Sets and Fuzzy Logic
- Explainable Artificial Intelligence (XAI)
- Multi-Criteria Decision Making
- Big Data and Business Intelligence
- Advanced Technologies in Various Fields
- Machine Learning and Data Classification
- Forecasting Techniques and Applications
- Computational Drug Discovery Methods
- Adversarial Robustness in Machine Learning
- Biomedical Text Mining and Ontologies
Zhejiang University
2022-2024
Mohamed bin Zayed University of Artificial Intelligence
2024
Zhejiang University of Science and Technology
2022-2024
Influenced by the great success of deep learning via cloud computing and rapid development edge chips, research in artificial intelligence (AI) has shifted to both paradigms, i.e., computing. In recent years, we have witnessed significant progress developing more advanced AI models on servers that surpass traditional owing model innovations (e.g., Transformers, Pretrained families), explosion training data soaring capabilities. However, computing, especially collaborative are still its...
In observational studies, confounder separation and balancing are the fundamental problems of treatment effect estimation. Most previous methods focused on addressing problem by treating all observed pre-treatment variables as confounders, ignoring separation. general, not confounders that refer to common causes outcome, some only contribute (i.e., instrumental variables) outcome adjustment variables). Balancing those non-confounders, including variables, would generate additional bias for...
Instrumental variables (IVs), sources of treatment randomization that are conditionally independent the outcome, play an important role in causal inference with unobserved confounders. However, existing IV-based counterfactual prediction methods need well-predefined IVs, while it is art rather than science to find valid IVs many real-world scenes. Moreover, predefined hand-made could be weak or erroneous by violating conditions IVs. These thorny facts hinder application methods. In this...
To address the growing size of AI model training data and lack a universal selection methodology-factors that significantly drive up costs -- this paper presents General Information Metrics Evaluation (GIME) method. GIME leverages general information metrics from Objective Theory (OIT), including volume, delay, scope, granularity, variety, duration, sampling rate, aggregation, coverage, distortion, mismatch to optimize dataset for purposes. Comprehensive experiments conducted across diverse...
Estimating the individuals' potential response to varying treatment doses is crucial for decision-making in areas such as precision medicine and management science. Most recent studies predict counterfactual outcomes by learning a covariate representation that independent of variable. However, independence constraints neglect much information useful prediction, especially when variables are continuous. To tackle above issue, this paper, we first theoretically demonstrate importance balancing...
Recent breakthroughs in artificial intelligence have driven a paradigm shift, where large language models (LLMs) with billions or trillions of parameters are trained on vast datasets, achieving unprecedented success across series tasks. However, despite these successes, LLMs still rely probabilistic modeling, which often captures spurious correlations rooted linguistic patterns and social stereotypes, rather than the true causal relationships between entities events. This limitation renders...
Learning individualized causal effect (ICE) plays a vital role in various fields of big data analysis, ranging from fine-grained policy evaluation to personalized treatment development. However, the presence unmeasured confounders increases difficulty estimating ICE real-world scenarios. A wide range methods have been proposed address with aid instrument variable (IV), which sources randomization. The performance these relies on well-predefined IVs that satisfy unconfounded instruments...
The fundamental problem in treatment effect estimation from observational data is confounder identification and balancing. Most of the previous methods realized balancing by treating all observed pre-treatment variables as confounders, ignoring further identifying confounders non-confounders. In general, not are that refer to common causes outcome, some only contribute outcome. Balancing those non-confounders, including instrumental adjustment variables, would generate additional bias for...
Causal inference is the process of using assumptions, study designs, and estimation strategies to draw conclusions about causal relationships between variables based on data. This allows researchers better understand underlying mechanisms at work in complex systems make more informed decisions. In many settings, we may not fully observe all confounders that affect both treatment outcome variables, complicating effects. To address this problem, a growing literature machine learning proposes...
The advent of the big data era brought new opportunities and challenges to draw treatment effect in fusion, that is, a mixed dataset collected from multiple sources (each source with an independent assignment mechanism). Due possibly omitted labels unmeasured confounders, traditional methods cannot estimate individual probability infer effectively. Therefore, we propose reconstruct label model it as Group Instrumental Variable (GIV) implement IV-based Regression for estimation. In this...
This paper focuses on developing Pareto-optimal estimation and policy learning to identify the most effective treatment that maximizes total reward from both short-term long-term effects, which might conflict with each other. For example, a higher dosage of medication increase speed patient's recovery (short-term) but could also result in severe side effects. Although recent works have investigated problems about or effects both, how trade-off between them achieve optimal remains an open...
Estimating the individuals' potential response to varying treatment doses is crucial for decision-making in areas such as precision medicine and management science. Most recent studies predict counterfactual outcomes by learning a covariate representation that independent of variable. However, independence constraints neglect much information useful prediction, especially when variables are continuous. To tackle above issue, this paper, we first theoretically demonstrate importance balancing...
Unobserved discrete data are ubiquitous in many scientific disciplines, and how to learn the causal structure of these latent variables is crucial for uncovering patterns. Most studies focus on linear variable model or impose strict constraints structures, which fail address cases involving non-linear relationships complex structures. To achieve this, we explore a tensor rank condition contingency tables an observed set $\mathbf{X}_p$, showing that determined by minimum support specific...
Heterogeneous treatment effect (HTE) estimation is vital for understanding the change of across individuals or subgroups. Most existing HTE methods focus on addressing selection bias induced by imbalanced distributions confounders between treated and control units, but ignore distribution shifts populations. Thereby, their applicability has been limited to in-distribution (ID) population, which shares a similar with training dataset. In real-world applications, where population are subject...
In causal inference, encouragement designs (EDs) are widely used to analyze effects, when randomized controlled trials (RCTs) impractical or compliance treatment cannot be perfectly enforced. Unlike RCTs, which directly allocate treatments, EDs randomly assign policies that positively motivate individuals engage in a specific treatment. These random encouragements act as instrumental variables (IVs), facilitating the identification of effects through leveraging exogenous perturbations...
Causal inference plays an important role in explanatory analysis and decision making across various fields like statistics, marketing, health care, education. Its main task is to estimate treatment effects make intervention policies. Traditionally, most of the previous works typically focus on binary setting that there only one for a unit adopt or not. However, practice, can be much more complex, encompassing multi-valued, continuous, bundle options. In this paper, we refer these as complex...
Learning directed acyclic graphs (DAGs) to identify causal relations underlying observational data is crucial but also poses significant challenges. Recently, topology-based methods have emerged as a two-step approach discovering DAGs by first learning the topological ordering of variables and then eliminating redundant edges, while ensuring that graph remains acyclic. However, one limitation these would generate numerous spurious edges require subsequent pruning. To overcome this...
Influenced by the great success of deep learning via cloud computing and rapid development edge chips, research in artificial intelligence (AI) has shifted to both paradigms, i.e., computing. In recent years, we have witnessed significant progress developing more advanced AI models on servers that surpass traditional owing model innovations (e.g., Transformers, Pretrained families), explosion training data soaring capabilities. However, computing, especially collaborative are still its...
The advent of the big data era brought new opportunities and challenges to draw treatment effect in fusion, that is, a mixed dataset collected from multiple sources (each source with an independent assignment mechanism). Due possibly omitted labels unmeasured confounders, traditional methods cannot estimate individual probability infer effectively. Therefore, we propose reconstruct label model it as Group Instrumental Variable (GIV) implement IV-based Regression for estimation. In this...
This paper studies the confounding effects from unmeasured confounders and imbalance of observed in IV regression aims at unbiased causal effect estimation. Recently, nonlinear estimators were proposed to allow for model both stages. However, may be imbalanced stage 2, which could still lead biased treatment estimation certain cases. To this end, we propose a Confounder Balanced Regression (CB-IV) algorithm jointly remove bias confounders. Theoretically, by redefining solving an inverse...