Requirements for Conceptual Representations of Explanations and How
            Reasoning Systems Can Serve Them

Horacek, Helmut

doi:10.18653/v1/w17-3703

Cited by 2 publications

(4 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Explaining the known gold answers for common-sense QA is an important research problem and is far from being solved (Rajani et al, 2019). Two major hurdles in solving this problem include (i) lack of any desiderata for what constitutes an explanation (Horacek, 2017) and (ii) unavailability of QA datasets comprising high quality human-annotated explanations.…”

Section: Cos Explanation (Rajani Et Al 2019)mentioning

confidence: 99%

“…For other QA tasks, such as common-sense QA, reading comprehension QA (RCQA), visual QA (VQA), grounding the definition of explanation is not so obvious (Horacek, 2017) and hence, they lack labeled data as well. In the case of RCQA and VQA (Ghosh et al, 2018), there have been attempts to explain the predicted answers.…”

Section: Free Flow Explanationmentioning

confidence: 99%

“…The broad idea behind explaining common-sense QA is to capture how humans would justify if a QA pair is presented to them. However, grounding a precise definition for this human justification is still hard due to subjectivity (Horacek, 2017). Furthermore, depending on the type of reasoning involved in the QA task, form and shape of an explanation may vary.…”

Section: Explanations For Commonsenseqamentioning

confidence: 99%

See 2 more Smart Citations

Explanations for CommonsenseQA: New Dataset and Models

Aggarwal¹,

Mandowara²,

Agrawal³

et al. 2021

Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Confer

View full text Add to dashboard Cite

CommonsenseQA (CQA) (Talmor et al., 2019) dataset was recently released to advance the research on common-sense question answering (QA) task. Whereas the prior work has mostly focused on proposing QA models for this dataset, our aim is to retrieve as well as generate explanation for a given (question, correct answer choice, incorrect answer choices) tuple from this dataset. Our explanation definition is based on certain desiderata, and translates an explanation into a set of positive and negative common-sense properties (aka facts) which not only explain the correct answer choice but also refute the incorrect ones. We human-annotate a first-ofits-kind dataset (called ECQA) of positive and negative properties, as well as free-flow explanations, for 11K QA pairs taken from the CQA dataset. We propose a latent representation based property retrieval model as well as a GPT-2 based property generation model with a novel two step fine-tuning procedure. We also propose a free-flow explanation generation model. Extensive experiments show that our retrieval model beats BM25 baseline by a relative gain of 100% in F 1 score, property generation model achieves a respectable F 1 score of 36.4, and free-flow generation model achieves a similarity score of 61.9, where last two scores are based on a human correlated semantic similarity metric.

show abstract

Section: Cos Explanation (Rajani Et Al 2019)mentioning

confidence: 99%

Section: Free Flow Explanationmentioning

confidence: 99%

Section: Explanations For Commonsenseqamentioning

confidence: 99%

See 1 more Smart Citation

Explanations for CommonsenseQA: New Dataset and Models

Aggarwal¹,

Mandowara²,

Agrawal³

et al. 2021

Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Confer

View full text Add to dashboard Cite

show abstract

“…In contrast to Machine Ethics, Machine Explainability aims at equipping complex and autonomous systems with means to make their decisions understandable to different groups of addressees (cf. [1,9,25,26,31]) enabling a sufficient amount of transparency and perspicuity for these systems. Doing so becomes more and more urgent: For instance, the software doping cases that surfaced in the context of the VW diesel emissions scandals made obvious that the behavior of complex systems can be very hard -if not practically impossible -to comprehend even for experts (cf.…”

Section: Introductionmentioning

confidence: 99%

Towards a Framework Combining Machine Ethics and Machine Explainability

Baum

Hermanns

Speith

2019

Electron. Proc. Theor. Comput. Sci.

View full text Add to dashboard Cite

We find ourselves surrounded by a rapidly increasing number of autonomous and semi-autonomous systems. Two grand challenges arise from this development: Machine Ethics and Machine Explainability. Machine Ethics, on the one hand, is concerned with behavioral constraints for systems, so that morally acceptable, restricted behavior results; Machine Explainability, on the other hand, enables systems to explain their actions and argue for their decisions in a way that human users can understand and justifiably trust them.In this paper, we try to motivate and work towards a framework combining Machine Ethics and Machine Explainability. Starting from a toy example, we detect various desiderata of such a framework and argue why they should and how they could be incorporated in autonomous systems. Our main idea is to apply a framework of formal argumentation theory both, for decision-making under ethically motivated constraints and for the task of generating useful explanations based on these constraints given only limited knowledge of the world. The result of our deliberations can be described as a first version of an ethically motivated, principle-governed framework combining Machine Ethics and Machine Explainability.

show abstract

Requirements for Conceptual Representations of Explanations and How Reasoning Systems Can Serve Them

Cited by 2 publications

References 6 publications

Explanations for CommonsenseQA: New Dataset and Models

Explanations for CommonsenseQA: New Dataset and Models

Towards a Framework Combining Machine Ethics and Machine Explainability

Contact Info

Product

Resources

About