Algorithmic Opacity: Making Algorithmic Processes Transparent through Abstraction Hierarchy

doi:10.1177/1541931218621046

Paper

Algorithmic Opacity: Making Algorithmic Processes Transparent through Abstraction Hierarchy

Published Sep 1, 2018 · Pragya Paudyal, B. L. William Wong

Proceedings of the Human Factors and Ergonomics Society Annual Meeting

UNKNOWN SJR score

12

Citations

1

Influential Citations

PDF

Semantic Scholar

Abstract

In this paper we introduce the problem of algorithmic opacity and the challenges it presents to ethical decision-making in criminal intelligence analysis. Machine learning algorithms have played important roles in the decision-making process over the past decades. Intelligence analysts are increasingly being presented with smart black box automation that use machine learning algorithms to find patterns or interesting and unusual occurrences in big data sets. Algorithmic opacity is the lack visibility of computational processes such that humans are not able to inspect its inner workings to ascertain for themselves how the results and conclusions were computed. This is a problem that leads to several ethical issues. In the VALCRI project, we developed an abstraction hierarchy and abstraction decomposition space to identify important functional relationships and system invariants in relation to ethical goals. Such explanatory relationships can be valuable for making algorithmic process transparent during the criminal intelligence analysis process.

Study Snapshot

Key takeawayAlgorithmic opacity in criminal intelligence analysis can be addressed through an abstraction hierarchy and decomposition space, enhancing ethical decision-making and reducing ethical concerns.

PopulationOlder adults (50-71 years)

Sample size24

MethodsObservational

OutcomesBody Mass Index projections

ResultsSocial networks mitigate obesity in older groups.

Sign up to use Study Snapshot

Consensus is limited without an account. Create an account or sign in to get more searches and use the Study Snapshot.

Create an account

Paper

Algorithmic Opacity: Making Algorithmic Processes Transparent through Abstraction Hierarchy

Sign up to use Study Snapshot

References

Towards A Rigorous Science of Interpretable Machine Learning

This paper proposes a taxonomy for rigorous evaluation of interpretable machine learning, aiming to define it, assess its need, and address open questions towards a more rigorous science of interpretable machine learning.

Algorithmic regulation and the global default: Shifting norms in Internet technology

This paper explores the governance mechanisms of content regulatory regimes and algorithms embedded in software, aiming to understand the normative frameworks embedded in technological systems.

Visualizing Large-scale and High-dimensional Data

LargeVis effectively visualizes large-scale and high-dimensional data in a low-dimensional space, reducing computational costs and improving efficiency and effectiveness compared to state-of-the-art methods.

The Threat of Algocracy: Reality, Resistance and Accommodation

Algorithmic governance poses a significant threat to the legitimacy of public decision-making processes, and resistance and accommodation are unlikely to effectively address this issue without compromising other values.

How the machine ‘thinks’: Understanding opacity in machine learning algorithms

Machine learning algorithms can be opacity-prone due to intentional secrecy, technical illiteracy, and scale requirements, requiring distinct solutions to prevent harm.

Explanation and understanding

People have skeletal understandings of the causal structure of the world, and can leverage these understandings to build explanations on the fly in real time, despite incomplete explanations.

Transparent Predictions

Transparency in predictive modeling is crucial for government efficiency, crowdsourcing, privacy, and autonomy, but must be balanced with concerns about government authority and stereotypes.

Citations

Towards Trustworthy AI: A Review of Ethical and Robust Large Language Models

Addressing ethical oversight, industry accountability, regulation, and public involvement can create a transparent and accountable AI ecosystem that benefits society while minimizing risks.

How a Daily Regimen of Operant Conditioning Might Explain the Power of the Search Engine Manipulation Effect (SEME)

The "search engine manipulation effect" (SEME) is a powerful list effect that influences undecided voters' opinions and voting preferences by shaping their opinions based on top search results.

Explanation and the Right to Explanation

The right to explanation in automated decision-making is challenging to extract from systems, and an outcomes-focused approach is recommended for pursuing the goods associated with explainability.

Analysis on the Impact of Recommender Systems on Consumer Decision: Making on China's Online Shopping Platforms

Recommender systems in online shopping platforms significantly influence consumers' decision-making and may influence their willingness-to-pay judgments, despite ethical challenges.

Formative Modeling of Foster Care Work: A Cognitive Work Analysis Approach

Cognitive Work Analysis (CWA) methodology can provide valuable insights into biases in foster care decision-making and help design context-sensitive decision aids.