Human-Guided Learning for Probabilistic Logic Models

Odom, Phillip; Natarajan, Sriraam

doi:10.3389/frobt.2018.00056

Cited by 8 publications

(14 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There are a number of interesting avenues for future work. Other interactive learning approaches such as coactive (Shivaswamy and others 2015), active imitation (Judah and others 2012), mixed-initiative interactive (Cakmak and others 2011) and guided probabilistic learning (Odom and Natarajan 2018) should be made explanatory. Making deep active learning (Gal and others 2017) explanatory is likely to improve upon the sample complexity of deep learning.…”

Section: Resultsmentioning

confidence: 99%

Explanatory Interactive Machine Learning

Teso

Kersting

2019

Proceedings of the 2019 AAAI/ACM Conference on AI, Ethics, and Society

146

148

View full text Add to dashboard Cite

Although interactive learning puts the user into the loop, the learner remains mostly a black box for the user. Understanding the reasons behind predictions and queries is important when assessing how the learner works and, in turn, trust. Consequently, we propose the novel framework of explanatory interactive learning where, in each step, the learner explains its query to the user, and the user interacts by both answering the query and correcting the explanation. We demonstrate that this can boost the predictive and explanatory powers of, and the trust into, the learned model, using text (e.g. SVMs) and image classification (e.g. neural networks) experiments as well as a user study.

show abstract

Section: Resultsmentioning

confidence: 99%

Explanatory Interactive Machine Learning

Teso

Kersting

2019

Proceedings of the 2019 AAAI/ACM Conference on AI, Ethics, and Society

146

148

View full text Add to dashboard Cite

show abstract

“…Since KiGB loss function is w.r.t the qualitative constraint and not tied with the gradient, it is easy to use this approach on any tree based learning methods like a decision tree, random-forests, AdaBoost, relational regression trees etc. As shown by Odom et al 2018 there exists close connection between qualitative constraints and preferences. In specific cases, preferences can be reduced to qualitative constraints and the KiGB framework can be leveraged.…”

Section: Extensionsmentioning

confidence: 89%

“…In specific cases, preferences can be reduced to qualitative constraints and the KiGB framework can be leveraged. Consider the example advice shown in (Odom and Natarajan 2018), if any car passes an agent on the right, then agent 3 https://starling.utdallas.edu/assets/pdfs/KokelAAAI20Sup.pdf should move into the right lane. This advice is represented as a preference rule (r = F, l+, l− ) with preferred label l+ = move right and avoid label l− = stay.…”

Section: Extensionsmentioning

confidence: 99%

“…Consequently, there has been a mild surge in developing human-allied systems that can solicit richer human inputs (Towell and Shavlik 1994;Kunapuli et al 2013;2010;Fung, Mangasarian, and Shavlik 2003). These inputs range from specifying domain knowledge in terms of decision-boundary constraints (Fung, Mangasarian, and Shavlik 2003), label preferences (Odom and Natarajan 2018), misclassification costs (Yang et al 2014), and qualitative constraints (Yang and Natarajan 2013) to name a few.…”

Section: Introductionmentioning

confidence: 99%

“…Specifically, we develop a unified approach that works for both classification and regression settings. While an earlier work attempted to address this problem by converting the qualitative constraints (QCs) to a preference framework (Odom and Natarajan 2018), it had two issues -first, the framework from QCs to preferences is not quite natural, second, and most importantly, the preference framework reweighed examples instead of using the QCs as constraints. From this perspective, QCs were used to alter the data distribution as against constraining the model as other probabilistic methods do (Altendorf, Restificar, and Dietterich 2005;Yang and Natarajan 2013).…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A Unified Framework for Knowledge Intensive Gradient Boosting: Leveraging Human Experts for Noisy Sparse Domains

Kokel

Odom

Yang

et al. 2020

AAAI

Self Cite

View full text Add to dashboard Cite

Incorporating richer human inputs including qualitative constraints such as monotonic and synergistic influences has long been adapted inside AI. Inspired by this, we consider the problem of using such influence statements in the successful gradient-boosting framework. We develop a unified framework for both classification and regression settings that can both effectively and efficiently incorporate such constraints to accelerate learning to a better model. Our results in a large number of standard domains and two particularly novel real-world domains demonstrate the superiority of using domain knowledge rather than treating the human as a mere labeler.

show abstract

Making deep neural networks right for the right scientific reasons by interacting with their explanations

et al. 2020

View full text Add to dashboard Cite

Deep neural networks have shown excellent performances in many real-world applications. Unfortunately, they may show "Clever Hans"-like behavior-making use of confounding factors within datasets-to achieve high performance.In this work, we introduce the novel learning setting of "explanatory interactive learning" (XIL) and illustrate its benefits on a plant phenotyping research task. XIL adds the scientist into the training loop, such that she interactively revises the original model via providing feedback on its explanations. Our experimental results demonstrate that XIL can help to avoid Clever Hans moments in machine learning and encourages (or discourages, if appropriate) trust into the underlying model. 1Imagine a plant phenotyping team attempting to characterize crop resistance to plant pathogens. The plant physiologist records a large amount of hyperspectral imaging data. Impressed by the results of deep learning in other scientific areas, she wants to establish similar results for phenotyping. Consequently, she asks a machine learning expert to apply deep learning to analyze the data. Luckily, the resulting predictive accuracy is very high. The plant physiologist, however, remains skeptical. The results are "too good, to be true". Checking the decision process of the deep model using explainable artificial intelligence (AI), the machine learning expert is flabbergasted to find that the learned deep model uses clues within the data that do not relate to the biological problem at hand, so-called confounding factors.The physiologist loses trust in AI and turns away from it, proclaiming it to be useless. This example encapsulates a critical issue of current explainable AI [1, 2]. Indeed, the seminal paper of Lapuschkin et al.[3] helps in "unmasking Clever Hans predictors and assessing what machines really learn", however, rather than proclaiming, as the plant physiologist might, that the machines have learned the right predictions for the wrong reasons and can therefore not be trusted, we here showcase that interactions between the learning system and the human user can correct the model, towards making the right predictions for the right reasons [4]. This may also increase the trust in machine learning models. Actually, trust lies at the foundation of major theories of interpersonal relationships in psychology [5, 6], and we argue that interaction and understandability are central to trust in learning machines. Surprisingly, the link between interacting, explaining, and building trust has been largely ignored by the machine learning literature. Existing approaches focus on passive learning only and do not consider the interaction between the user and the learner [7,8,9], whereas, interactive learning frameworks such as active [10] and coactive learning [11] do not consider the issue of trust. In active learning, for instance, the model presents unlabeled instances to a user, and in exchange, obtains their label. This is completely opaque-the user is oblivious to the models beliefs and reasons for predictions...

show abstract

Human-Guided Learning for Probabilistic Logic Models

Cited by 8 publications

References 41 publications

Explanatory Interactive Machine Learning

Explanatory Interactive Machine Learning

A Unified Framework for Knowledge Intensive Gradient Boosting: Leveraging Human Experts for Noisy Sparse Domains

Making deep neural networks right for the right scientific reasons by interacting with their explanations

Contact Info

Product

Resources

About