More Than Accuracy: Towards Trustworthy Machine Learning Interfaces for Object Recognition

Heuer, Hendrik; Breiter, Andreas

doi:10.1145/3340631.3394873

Cited by 8 publications

(4 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, a pilot study in October 2020 provided evidence that explanations like "The average number of words per sentence is low", the "Usage of perceptual words related to hear is low" and "The amount of words related to people is high" are not perceived as helpful by users. Even educated participants had trouble understanding explanations of ML-based systems, a problem that I have previously reported on in the context of ML-based curation systems [3] and object recognition systems [6].…”

Section: How People Can Be Best Supportedmentioning

confidence: 98%

Helping People Deal With Disinformation -- A Socio-Technical Perspective

Heuer

2021

Preprint

Self Cite

View full text Add to dashboard Cite

At the latest since the advent of the Internet, disinformation and conspiracy theories have become ubiquitous. Recent examples like QAnon and Pizzagate prove that false information can lead to real violence. Sadly, this link between misinformation and violence has a long history. Anti-semitic conspiracy theories played a central role in the Shoa. In the context of the COVID-19 pandemic, the World Health Organization (WHO) warned that "misinformation costs lives" [13]. The WHO argued that mitigating the harm from misinformation is necessary to manage the COVID-19 infodemic. The recent increase in misinformation in a so-called Post-Truth Era can be linked to societal mega-trends such as a decline in social capital, growing economic inequality, increased polarization, declining trust in science, and an increasingly fractionated media landscape [9].My work is focused on disinformation. Following Claire Wardle [2], I operationalize disinformation as fabricated or deliberately manipulated content like conspiracy theories or rumors. As such, disinformation is a special case of misinformation. While disinformation is connected to an intent to harm, misinformation also includes unintentional mistakes. I believe that disinformation is a challenging, multifaceted phenomenon that requires an appropriate sociotechnical response. In my work, I research 1. why people believe in disinformation, 2. how people can be best supported in recognizing disinformation, and 3. what the potentials and risks of different tools designed to fight disinformation are.My work on disinformation is informed by my background in human-computer interaction and machine learning.The workshop would be a great opportunity for me to discuss the ethical implications of the disinformation detection solutions that I am developing. I would love to discuss ways of enabling fast and effective disinformation detection while ensuring that freedom of speech is protected. MACHINE LEARNING-BASED CURATION SYSTEMSTo understand disinformation in the contemporary media climate, one has to understand social media and the machine learning-based curation systems used on social media. My doctoral thesis provides a socio-technical perspective on users and machine learning-based curation systems [3]. The thesis presents actionable insights on how ML-based curation systems can and should be explained and audited. Motivated by the role that ML-based curation systems play in the dissemination of disinformation, I examined the user beliefs around such systems in detail. In a recent CSCW paper, I, together with my collaborators, examined how users without a technical background, who regularly Permission to make digital or hard copies of part or all of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for third-party components of this work must be honored. For all other uses, contact the owner/author(s).

show abstract

Section: How People Can Be Best Supportedmentioning

confidence: 98%

Helping People Deal With Disinformation -- A Socio-Technical Perspective

Heuer

2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…The emergence of complex, opaque, and invisible algorithms that learn from data motivated a variety of investigations, including: algorithm awareness, clarity, variance, and bias [94]. Algorithmic bias for instance, whether it occurs in an unintentional or intentional manner, is found to severely limit the performance of an AI model.…”

Section: Recommendations and The Future Of Ai Assurancementioning

confidence: 99%

A survey on artificial intelligence assurance

2021

View full text Add to dashboard Cite

Artificial Intelligence (AI) algorithms are increasingly providing decision making and operational support across multiple domains. AI includes a wide (and growing) library of algorithms that could be applied for different problems. One important notion for the adoption of AI algorithms into operational decision processes is the concept of assurance. The literature on assurance, unfortunately, conceals its outcomes within a tangled landscape of conflicting approaches, driven by contradicting motivations, assumptions, and intuitions. Accordingly, albeit a rising and novel area, this manuscript provides a systematic review of research works that are relevant to AI assurance, between years 1985 and 2021, and aims to provide a structured alternative to the landscape. A new AI assurance definition is adopted and presented, and assurance methods are contrasted and tabulated. Additionally, a ten-metric scoring system is developed and introduced to evaluate and compare existing methods. Lastly, in this manuscript, we provide foundational insights, discussions, future directions, a roadmap, and applicable recommendations for the development and deployment of AI assurance.

show abstract

“…We included the system predictions as intuitive explanations because they present the predictions in a format that is similar to how news recommendations are encountered by users [35,4,32]. We also presented the participants with the three most important Performance Metrics for ML systems: accuracy, precision, and recall [17,22]. Accuracy is defined as the percentage of correctly predicted news, i.e.…”

Section: Explanations For Ml-based Curation Systemmentioning

confidence: 99%

The Explanatory Gap in Algorithmic News Curation

Heuer

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Considering the large amount of available content, social media platforms increasingly employ machine learning (ML) systems to curate news. This paper examines how well different explanations help expert users understand why certain news stories are recommended to them. The expert users were journalists, who are trained to judge the relevance of news. Surprisingly, none of the explanations are perceived as helpful. Our investigation provides a first indication of a gap between what is available to explain ML-based curation systems and what users need to understand such systems. We call this the Explanatory Gap in Machine Learning-based Curation Systems.

show abstract

More Than Accuracy: Towards Trustworthy Machine Learning Interfaces for Object Recognition

Cited by 8 publications

References 35 publications

Helping People Deal With Disinformation -- A Socio-Technical Perspective

Helping People Deal With Disinformation -- A Socio-Technical Perspective

A survey on artificial intelligence assurance

The Explanatory Gap in Algorithmic News Curation

Contact Info

Product

Resources

About