Explainable digital forensics AI: Towards mitigating distrust in AI-based digital forensics analysis using interpretable models

Solanke, Abiodun A.

doi:10.1016/j.fsidi.2022.301403

Cited by 23 publications

(11 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Feature extraction detects and separates fundamental portions of digital images in object detection and classification. Extracting ideal features in a reduced order that can reflect the most relevant content of the images for image detection is still a challenging problem [13]. Very little research has paid attention to this problem.…”

Section: Methodsmentioning

confidence: 99%

Decoding Innocence: Advancing Forensic Facial Discrimination through Comparative Analysis of Conventional CNN and Advanced Architectures

Ahmed,

Abdullah

2024

JOIV : Int. J. Inform. Visualization

View full text Add to dashboard Cite

The field of Digital Image Forensics (DIF) faces a critical issue in accurately identifying children in digital images, notably in cases involving the proliferation of child sexual abuse content. Existing techniques face hurdles due to model architecture limitations, dataset suitability concerns, and classification imbalance, impeding their ability to recognize children to deter pornographic images. Addressing this challenge, this study introduces Implicit Feature Extraction (IFE), a specialized approach for distinguishing child and adult images in object detection. Leveraging Convolutional Neural Networks (CNNs), the IFE method automates the extraction of discriminative facial features, surpassing the constraints of Explicit Feature Extraction (EFE) methods, which achieve an accuracy of around 70%. The research focuses on three core objectives introducing IFE for detailed face feature detection in DIF's child and adult image identification, implementing IFE with CNNs to enhance image classification, and conducting a thorough evaluation of the proposed technique's performance using key metrics like accuracy and balanced classification results and comparing the result with a basic CNN model’s performance. This research's significance lies in its notable contributions to digital image forensics, particularly in combatting child exploitation. The fusion of IFE with CNNs showcases 92% accuracy in distinguishing child and adult images, promising advancements with practical implications in child protection and forensic investigations. The comprehensive evaluation using the UTKFace dataset underscores the proposed technique's efficacy, marking a substantial improvement in child image identification within digital image forensics.

show abstract

Section: Methodsmentioning

confidence: 99%

Decoding Innocence: Advancing Forensic Facial Discrimination through Comparative Analysis of Conventional CNN and Advanced Architectures

Ahmed,

Abdullah

2024

JOIV : Int. J. Inform. Visualization

View full text Add to dashboard Cite

show abstract

“…In addition, providers should disclose relevant information to potential customers ( Porto, 2021 ). Interpretable models are used to construct interpretable AI-based digital forensics ( Solanke, 2022 ).…”

Section: Progress Of Cognitive Methods Applied To Judicial Modelmentioning

confidence: 99%

The application of cognitive neuroscience to judicial models: recent progress and trends

Zhang,

Zhang

2023

Front. Neurosci.

View full text Add to dashboard Cite

Legal prediction presents one of the most significant challenges when applying artificial intelligence (AI) to the legal field. The legal system is a complex adaptive system characterized by the ambiguity of legal language and the diversity of value functions. The imprecision and procedural knowledge inherent in law makes judicial issues difficult to be expressed in a computer symbol system. Current semantic processing and machine learning technologies cannot fully capture the complex nature of legal relations, thereby raising doubts about the accuracy of legal predictions and reliability of judicial models. Cognitive computing, designed to emulate human brain functions and aid in enhancing decision-making processes, offers a better understanding of legal data and the processes of legal reasoning. This paper discusses the advancements made in cognitive methods applied to legal concept learning, semantic extraction, judicial data processing, legal reasoning, understanding of judicial bias, and the interpretability of judicial models. The integration of cognitive neuroscience with law has facilitated several constructive attempts, indicating that the evolution of cognitive law could be the next frontier in the intersection of AI and legal practice.

show abstract

“…Considering the problems of algorithmic governance in the public sector, Maciej Kuziemski and Gianluca Misuraca write that 'agenda setting bottlenecks are further perpetuated by misaligned incentives, goals and measures: public sector's duties towards the citizens are at odds with those of the profit maximizing private sector' . 25 Despite the fact that profit maximisation as a goal of the private sector is nothing fundamentally new, the growth in the number of technological solutions offered by private developers and used in the public sector makes us look at it more closely. The lack of ethical basis for such decisions and the absence of stable standards for them can leadand already have led -to growing inequality and undermining the value foundations of justice.…”

Section: Judicial Independence Under the Influence Of Digitalisationmentioning

confidence: 99%

“…This includes, for example, some concerns 'about the legality of digital evidence or machine-generated conclusions, particularly given that these decisions can differ for the same scientific evidence, just as they do with human experts' . 33 This is also reflected in concerns of a different kind, which 'is beyond the level of "access" to justice and is inherently relevant to the quality of judicial services. Essentially, the question is whether the use of artificial intelligence technology and algorithms in court systems could produce "just" outcomes' .…”

Section: The Algorithmisation Of Decision-makingmentioning

confidence: 99%

Algorithms in the Courts: Is There Any Room for a Rule of Law?

2022

AJEE

View full text Add to dashboard Cite

The rule of law is one of the fundamental pillars, along with human rights and democracy, which are affected by digitalisation today. Digital technologies used for the victory of populism, the manipulation of opinions, attacks on the independence of judges, and the general instrumentalisation of the law contribute significantly to the onset of negative consequences for the rule of law. Particularly dangerous are the far-reaching consequences of the algorithmisation of decision-making, including judicial decisions. The theoretical line of this research is based on the axiological method since the rule of law, democracy, and human rights are not only the foundations of legal order, but also values recognised in many societies and supported at the individual level. The study also relied on the phenomenological method in terms of assessing the experience of being influenced by digital technologies in public and private life. The practical line of research is based on the analysis of cases of the European Court of Human Rights and the Court of Justice to illustrate the changes in jurisprudence influenced by digitalisation. This article argues that the potential weakening of the rule of law could be related to the impact of certain technologies itself, and to their impact on certain values and foundations which is significantly aggravated. Judicial independence is affected since the judges are involved in digital interactions and are influenced by technologies along personal and public lines. That technologies often belong private sector but are perceived as neutral and infallible, which is highly predictive of court decisions. This leads to a distortion of the essence of legal certainty and a shift of trust from the courts to certain technologies and their creators. The possibility of algorithmic decision-making raises the question of whether the results will be fairer, or at least as fair, as those handed down by human judges. This entails two problems, the first of which is related to the task of interpreting the law and the second of which involves the need to explain decisions. Algorithms, often perceived as reliable, are not really capable of interpreting the law, and their ability to provide proper explanations for decisions or understand context and social practices is questionable. Even partial reliance on algorithms should be limited, given the growing inability to draw a line between the human and algorithmic roles in decision-making and determine who should be responsible for the decision and to what extent.

show abstract

Explainable digital forensics AI: Towards mitigating distrust in AI-based digital forensics analysis using interpretable models

Cited by 23 publications

References 37 publications

Decoding Innocence: Advancing Forensic Facial Discrimination through Comparative Analysis of Conventional CNN and Advanced Architectures

Decoding Innocence: Advancing Forensic Facial Discrimination through Comparative Analysis of Conventional CNN and Advanced Architectures

The application of cognitive neuroscience to judicial models: recent progress and trends

Algorithms in the Courts: Is There Any Room for a Rule of Law?

Contact Info

Product

Resources

About