Causal inference and the data-fusion problem

Great care is taken in epidemiologic studies to ensure the internal validity of causal effect estimates; however, external validity has received considerably less attention. When the study sample is not a random sample of the target population, the sample average treatment effect, even if internally valid, cannot usually be expected to equal the average treatment effect in the target population. The utility of an effect estimate for planning purposes and decision making will depend on the degree of departure from the true causal effect in the target population due to problems with both internal and external validity. Herein, we review concepts from recent literature on generalizability, one facet of external validity, using the potential outcomes framework. Identification conditions sufficient for external validity closely parallel identification conditions for internal validity, namely: conditional exchangeability; positivity; the same distributions of the versions of treatment; no interference; and no measurement error. We also require correct model specification. Under these conditions, we discuss how a version of direct standardization (the g-formula, adjustment formula, or transport formula) or inverse probability weighting can be used to generalize a causal effect from a study sample to a well-defined target population, and demonstrate their application in an illustrative example.

show abstract

“…Methods for handling both selection and transportability problems are surveyed by Bareinboim and colleagues. 3,4,44 …”

Section: Transportabilitymentioning

confidence: 99%

Generalizing Study Results

et al. 2017

View full text Add to dashboard Cite

show abstract

“…Causal inference deals with the problem of inferring the effect of actions (target) from a combination of a causal model (to be defined) and heterogeneous sources of data (source) [Pearl, 2000;Bareinboim and Pearl, 2016]. One of the fundamental challenges in the field is to determine whether a causal effect can be inferred from the observational (nonexperimental) distribution when important variables in the problem may be unmeasured (also called unobserved confounders, or UCs).…”

Section: Introductionmentioning

confidence: 99%

“…Despite its success in identifying the effect of actions from heterogeneous data in compelling settings across the sciences [Bareinboim and Pearl, 2016], causal inference techniques have rarely been used to assist the transfer of knowledge in interactive domains. [Mehta et al, 2008;] assumed a causal model for the underlying task and performed the transfer of probabilistic knowledge leveraging the invariance encoded in the causal model.…”

Section: Introductionmentioning

confidence: 99%

Transfer Learning in Multi-Armed Bandits: A Causal Approach

Zhang

Bareinboim

2017

Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence

Self Cite

View full text Add to dashboard Cite

Reinforcement learning (RL) agents have been deployed in complex environments where interactions are costly, and learning is usually slow. One prominent task in these settings is to reuse interactions performed by other agents to accelerate the learning process. Causal inference provides a family of methods to infer the effects of actions from a combination of data and qualitative assumptions about the underlying environment. Despite its success of transferring invariant knowledge across domains in the empirical sciences, causal inference has not been fully realized in the context of transfer learning in interactive domains. In this paper, we use causal inference as a basis to support a principled and more robust transfer of knowledge in RL settings. In particular, we tackle the problem of transferring knowledge across bandit agents in settings where causal effects cannot be identified by do-calculus [Pearl, 2000] and standard learning techniques. Our new identification strategy combines two steps -first, deriving bounds over the arms distribution based on structural knowledge; second, incorporating these bounds in a dynamic allocation procedure so as to guide the search towards more promising actions. We formally prove that our strategy dominates previously known algorithms and achieves orders of magnitude faster convergence rates than these algorithms. Finally, we perform simulations and empirically demonstrate that our strategy is consistently more efficient than the current (non-causal) state-of-the-art methods.

show abstract

“…They help greatly to make a decision. The application fields of the data fusion are varied and diverse: Medical imaging (Magtibay et al, 2016;Adali et al, 2015), economy (Barenboim and Pearl, 2016), information theory (Gagolewski, 2016), image processing (Paris and Bruzzone, 2015), etc. In remote sensing where the nature and the resolution of sensors are various and different, methods of image fusion implement several types of images: Panchromatic (PAN), Multispectral (MS), Hyperspectral (HS) and Radar (SAR).…”

Section: Introductionmentioning

confidence: 99%

Application of Morphological Component Analysis to Optical Image Fusion

Loum¹,

Armand²,

Ghislain³

et al. 2017

American Journal of Applied Sciences

View full text Add to dashboard Cite

Abstract:The image fusion technique is widely used in remote sensing. Its purpose is to provide comprehensive information without arte facts by combining the partial information from different source images. In this study, we propose a new model of images fusion with very high spatial resolution. We use the separation capacities of the Morphological Component Analysis (MCA) to extract the smooth and texture components of our images. These morphological components are then fused separately using the decomposition in the Laplacian pyramids for the smooth part and bivariate Hahn polynomials for texture part. Finally the image fusion is obtained through linear combination of merged smooth and texture components. The experiments carried out on IKONOS, LANDSAT and Quick Bird remote sensing images show the good performances of our method which has been compared to conventional methods. The performances obtained in our experiments are characterized by a small global metric such as ERGAS equals to 3.88 for IKONOS image and 3.65 for QuickBird image compared to 8.70 for IKONOS image and 6.97 for QuickBird for conventional HIS algorithms. We also have a mean loss of 15% for spectral information com pare to those of the conventional methods which revolve around 25%. The degradation of spatial information in order of 17% in contrast to conventional HIS algorithms which oscillate around 21%.

show abstract

Causal inference and the data-fusion problem

Cited by 497 publications

References 41 publications

Generalizing Study Results

Generalizing Study Results

Transfer Learning in Multi-Armed Bandits: A Causal Approach

Application of Morphological Component Analysis to Optical Image Fusion

Contact Info

Product

Resources

About