Humans In Loop Research Articles

Human in Loop Machine Learning: A Paradigm Shift

This paper explores the transformative concept of Human-in-Loop (HIL) Machine Learning, which integrates human expertise into the machine learning process to enhance data quality, model accuracy, and ethical decision-making. Human interaction is added to traditional machine learning steps such data collection, preprocessing, model training, evaluation, and deployment through continuous feedback loops, data annotation, and model change. This integration makes use of human intuition and experience to enhance algorithm performance and model interpretability, hence mitigating the drawbacks of entirely automated approaches. HIL Machine Learning allows AI systems to adjust to changing obstacles and makes more accurate forecasts by promoting human-machine interaction. This study emphasizes HIL Machine Learning as a reliable method for successfully handling dynamic, complicated problems across a range of areas.

ChatGPT-based biological and psychological data imputation

Missing data are a common problem for large cohort or longitudinal research and have been handled through data imputation. Based on simplified models such as linear or nonlinear interpolations, current imputation methods may not be accurate for real-life data such as biological and behavioral data. The purpose of this work was to explore the capability of ChatGPT, a powerful Large Language Model (LLM) developed by OpenAI, for biological and psychological data imputation. We tested the feasibility using data from the Human Connectome Project. Performance was evaluated by comparing the imputed data against known ground truth (GT) and measured with metrics like Pearson correlation coefficient (r), relative accuracy (MP), and mean absolute error (MAE). Comparative analyses with traditional imputation techniques are also conducted to demonstrate the superior efficacy of the ChatGPT as a data imputer. In summary, through customized data-to-text prompting engineering, ChatGPT can successfully capture intricate patterns and dependencies within biological data, resulting in precise imputations. Fine-tuning ChatGPT with domain-specific biological vocabulary with human in-loop as an interpreter enhances the accuracy and relevance of the imputations.

Ultralow-Latency Textile Sensors for Wearable Interfaces with a Human-in-Loop Sensing Approach.

The evolution of wearable technologies has led to the development of novel types of sensors customized for a wide range of applications. Wearable sensors need to possess a low form factor and be ergonomic, causing minimal impediment of the user's natural movement. Various principles have been explored to meet these requirements, ranging from optical, magnetic, resistive flex sensing to 3D printed sensors and liquid metals such as those using eutectic gallium-indium. However, manufacturing techniques for most current wearable sensors tend to be complex and difficult to scale. Challenges also exist in achieving high sensitivity with noise resistance and robustness to false detections, especially in capacitive sensors. In this research, a novel ultralow-latency soft tactile and pressure sensor developed using off-the-shelf e-textiles is proposed, which overcomes some of these limitations. The sensor does not use any specialized equipment or materials for manufacture. A human-in-loop (HIL) sensing technique is demonstrated, which provides high sensitivity, high sensing bandwidth, as well as ultralow latency, which makes it ideal as a wearable input device. In addition, the HIL method provides other advantages such as high noise rejection and resistance to accidental triggers that could be caused by other humans or environmental factors owing to its high signal to noise ratio. Finally, two applications-a wearable keyboard and gaming input device-were demonstrated using these sensors.

Liable, but Not in Control? Ensuring Meaningful Human Agency in Automated Decision‐Making Systems

Automated decision making is becoming the norm across large parts of society, which raises interesting liability challenges when human control over technical systems becomes increasingly limited. This article defines “quasi‐automation” as inclusion of humans as a basic rubber‐stamping mechanism in an otherwise completely automated decision‐making system. Three cases of quasi‐automation are examined, where human agency in decision making is currently debatable: self‐driving cars, border searches based on passenger name records, and content moderation on social media. While there are specific regulatory mechanisms for purely automated decision making, these regulatory mechanisms do not apply if human beings are (rubber‐stamping) automated decisions. More broadly, most regulatory mechanisms follow a pattern of binary liability in attempting to regulate human or machine agency, rather than looking to regulate both. This results in regulatory gray areas where the regulatory mechanisms do not apply, harming human rights by preventing meaningful liability for socio‐technical decision making. The article concludes by proposing criteria to ensure meaningful agency when humans are included in automated decision‐making systems, and relates this to the ongoing debate on enabling human rights in Internet infrastructure.

Relevance in cyber‐physical systems with humans in the loop

SummaryIn cyber‐physical systems such as intelligent lighting, the system responds autonomously to observed changes in the environment. In such systems, more than one output may be acceptable for a given input scenario. This type of relationship between the input and output makes it difficult to analyze machine learning algorithms using commonly used performance metrics such as classification accuracy (CA). CA only measures whether a predicted output is right or not, whereas it is more important to determine whether the predicted output is relevant for the given context or not. In this direction, we introduce a new metric, the relevance score (RS) that is effective for the class of applications where user perception leads to non‐deterministic input–output relationships. RS determines the extent by which a predicted output is relevant to the user's context and behaviors, taking into account the variability and bias that come with human perception factors. We assess the performance of a number of machine learning algorithms, using different datasets, including data from an intelligent lighting pilot. We find that using RS instead of CA is appropriate to analyze the performance of conventional machine learning algorithms, particularly for the class of non‐deterministic multiple‐output problems. Our method may be applied to other scenarios in which cyber‐physical systems involve humans in the control loop. Copyright © 2016 John Wiley & Sons, Ltd.