Search

In this paper, we investigate the counter-forensic effects of the forthcoming JPEG AI standard based on neural image compression, focusing on two critical areas: deepfake image detection and image splicing localization. Neural image compression leverages advanced neural network algorithms to achieve higher compression rates while maintaining image quality. However, it introduces artifacts that closely resemble those generated by image synthesis techniques and image splicing pipelines, complicating the work of researchers when discriminating pristine from manipulated content. We comprehensively analyze JPEG AI's counter-forensic effects through extensive experiments on several state-of-the-art detectors and datasets. Our results demonstrate that an increase in false alarms impairs the performance of leading forensic detectors when analyzing genuine content processed through JPEG AI. By exposing the vulnerabilities of the available forensic tools we aim to raise the urgent need for multimedia forensics researchers to include JPEG AI images in their experimental setups and develop robust forensic techniques to distinguish between neural compression artifacts and actual manipulations.

Labor plays a major, albeit largely unrecognized role in the development of artificial intelligence. Machine learning algorithms are predicated on data-intensive processes that rely on humans to execute repetitive and difficult-to-automate, but no less essential, tasks such as labeling images, sorting items in lists, recording voice samples, and transcribing audio files. Online platforms and networks of subcontractors recruit data workers to execute such tasks in the shadow of AI production, often in lower-income countries with long-standing traditions of informality and lessregulated labor markets. This study unveils the resulting complexities by comparing the working conditions and the profiles of data workers in Venezuela, Brazil, Madagascar, and as an example of a richer country, France. By leveraging original data collected over the years 2018-2023 via a mixed-method design, we highlight how the cross-country supply chains that link data workers to core AI production sites are reminiscent of colonial relationships, maintain historical economic dependencies, and generate inequalities that compound with those inherited from the past. The results also point to the importance of less-researched, non-English speaking countries to understand key features of the production of AI solutions at planetary scale., Comment: Jack Qiu, Shinjoung Yeo, Richard Maxwell. The Handbook of Digital Labor, Wiley Blackwell, In press, ISBN10: 1119981808

When dealing with multimedia data, source attribution is a key challenge from a forensic perspective. This task aims to determine how a given content was captured, providing valuable insights for various applications, including legal proceedings and integrity investigations. The source attribution problem has been addressed in different domains, from identifying the camera model used to capture specific photographs to detecting the synthetic speech generator or microphone model used to create or record given audio tracks. Recent advancements in this area rely heavily on machine learning and data-driven techniques, which often outperform traditional signal processing-based methods. However, a drawback of these systems is their need for large volumes of training data, which must reflect the latest technological trends to produce accurate and reliable predictions. This presents a significant challenge, as the rapid pace of technological progress makes it difficult to maintain datasets that are up-to-date with real-world conditions. For instance, in the task of smartphone model identification from audio recordings, the available datasets are often outdated or acquired inconsistently, making it difficult to develop solutions that are valid beyond a research environment. In this paper we present POLIPHONE, a dataset for smartphone model identification from audio recordings. It includes data from 20 recent smartphones recorded in a controlled environment to ensure reproducibility and scalability for future research. The released tracks contain audio data from various domains (i.e., speech, music, environmental sounds), making the corpus versatile and applicable to a wide range of use cases. We also present numerous experiments to benchmark the proposed dataset using a state-of-the-art classifier for smartphone model identification from audio recordings., Comment: Submitted to IEEE Access

While much has been written about the problem of information overload in news and social media, little attention has been paid to its consequence in science. Scientific literature, however, has witnessed decades of exponential growth, to the point that the publications of the last twenty years now constitute 60% of all academic literature. This information overload is not without consequence. Our analysis reveals that, unlike other cultural products, scientific publications face unique challenges: the decreasing proportion of papers capturing large shares of researchers' attention and the slow turnover of influential papers lead to a disproportionate prominence of established works, resulting in stagnation and aging of scientific canons. To determine whether scientific hypergrowth is responsible for such ``gerontocratization of science'', we propose a generative model of paper citations based on random discovery and cumulative advantage, with a varying number of new papers each year. Our findings show that, as exponential growth intensifies, gerontocratization appears and becomes increasingly pronounced. Recognizing and understanding this mechanism is hence essential for developing targeted strategies to counteract this trend and promote a balanced and healthy renewal of scientific canons., Comment: 19 pages, 7 figures

In speech deepfake detection, one of the critical aspects is developing detectors able to generalize on unseen data and distinguish fake signals across different datasets. Common approaches to this challenge involve incorporating diverse data into the training process or fine-tuning models on unseen datasets. However, these solutions can be computationally demanding and may lead to the loss of knowledge acquired from previously learned data. Continual learning techniques offer a potential solution to this problem, allowing the models to learn from unseen data without losing what they have already learned. Still, the optimal way to apply these algorithms for speech deepfake detection remains unclear, and we do not know which is the best way to apply these algorithms to the developed models. In this paper we address this aspect and investigate whether, when retraining a speech deepfake detector, it is more effective to apply continual learning across the entire model or to update only some of its layers while freezing others. Our findings, validated across multiple models, indicate that the most effective approach among the analyzed ones is to update only the weights of the initial layers, which are responsible for processing the input features of the detector., Comment: Submitted to ICASSP 2025

Speech deepfakes pose a significant threat to personal security and content authenticity. Several detectors have been proposed in the literature, and one of the primary challenges these systems have to face is the generalization over unseen data to identify fake signals across a wide range of datasets. In this paper, we introduce a novel approach for enhancing speech deepfake detection performance using a Mixture of Experts architecture. The Mixture of Experts framework is well-suited for the speech deepfake detection task due to its ability to specialize in different input types and handle data variability efficiently. This approach offers superior generalization and adaptability to unseen data compared to traditional single models or ensemble methods. Additionally, its modular structure supports scalable updates, making it more flexible in managing the evolving complexity of deepfake techniques while maintaining high detection accuracy. We propose an efficient, lightweight gating mechanism to dynamically assign expert weights for each input, optimizing detection performance. Experimental results across multiple datasets demonstrate the effectiveness and potential of our proposed approach., Comment: Submitted to ICASSP 2025

AI generates both enthusiasm and disillusionment, with promises that often go unfulfilled. It is therefore not surprising that human labor, which is its fundamental component, is also subject to these same deceptions. The development of "smart technologies" depends, at different stages, on a multitude of precarious, underpaid and invisible workers, who, dispersed globally, carry out repetitive, fragmented activities, paid per task and completed in a few seconds. These are workers who label data to train algorithms, through tasks that require the intuitive, creative and cognitive abilities of human beings, such as categorizing images, classifying advertisements, transcribing audio and video, evaluating advertisements, moderating content on social media, labeling human anatomical points of interest, digitizing documents, etc. This form of work is often referred to as "microwork". Our contribution, which documents the conditions of microwork in Brazil and offers portraits of the workers, is a step in the wider effort to overcome the current state of invisibilization. It opens up avenues for future research, with the aim of better characterizing this new form of work, tracing its changes over time in relation to the dynamics of globalization and, ideally, identifying levers for action and transitions., Comment: in Portuguese language

Text-To-Music (TTM) models have recently revolutionized the automatic music generation research field. Specifically, by reaching superior performances to all previous state-of-the-art models and by lowering the technical proficiency needed to use them. Due to these reasons, they have readily started to be adopted for commercial uses and music production practices. This widespread diffusion of TTMs poses several concerns regarding copyright violation and rightful attribution, posing the need of serious consideration of them by the audio forensics community. In this paper, we tackle the problem of detection and attribution of TTM-generated data. We propose a dataset, FakeMusicCaps that contains several versions of the music-caption pairs dataset MusicCaps re-generated via several state-of-the-art TTM techniques. We evaluate the proposed dataset by performing initial experiments regarding the detection and attribution of TTM-generated audio.

Speech deepfake detection has recently gained significant attention within the multimedia forensics community. Related issues have also been explored, such as the identification of partially fake signals, i.e., tracks that include both real and fake speech segments. However, generating high-quality spliced audio is not as straightforward as it may appear. Spliced signals are typically created through basic signal concatenation. This process could introduce noticeable artifacts that can make the generated data easier to detect. We analyze spliced audio tracks resulting from signal concatenation, investigate their artifacts and assess whether such artifacts introduce any bias in existing datasets. Our findings reveal that by analyzing splicing artifacts, we can achieve a detection EER of 6.16% and 7.36% on PartialSpoof and HAD datasets, respectively, without needing to train any detector. These results underscore the complexities of generating reliable spliced audio data and lead to discussions that can help improve future research in this area., Comment: Accepted at ASVspoof 5 Workshop (Interspeech2024 Satellite)

AI-generated synthetic media, also called Deepfakes, have significantly influenced so many domains, from entertainment to cybersecurity. Generative Adversarial Networks (GANs) and Diffusion Models (DMs) are the main frameworks used to create Deepfakes, producing highly realistic yet fabricated content. While these technologies open up new creative possibilities, they also bring substantial ethical and security risks due to their potential misuse. The rise of such advanced media has led to the development of a cognitive bias known as Impostor Bias, where individuals doubt the authenticity of multimedia due to the awareness of AI's capabilities. As a result, Deepfake detection has become a vital area of research, focusing on identifying subtle inconsistencies and artifacts with machine learning techniques, especially Convolutional Neural Networks (CNNs). Research in forensic Deepfake technology encompasses five main areas: detection, attribution and recognition, passive authentication, detection in realistic scenarios, and active authentication. This paper reviews the primary algorithms that address these challenges, examining their advantages, limitations, and future prospects.

In recent years, methods for producing highly realistic synthetic images have significantly advanced, allowing the creation of high-quality images from text prompts that describe the desired content. Even more impressively, Stable Diffusion (SD) models now provide users with the option of creating synthetic images in an image-to-image translation fashion, modifying images in the latent space of advanced autoencoders. This striking evolution, however, brings an alarming consequence: it is possible to pass an image through SD autoencoders to reproduce a synthetic copy of the image with high realism and almost no visual artifacts. This process, known as SD image laundering, can transform real images into lookalike synthetic ones and risks complicating forensic analysis for content authenticity verification. Our paper investigates the forensic implications of image laundering, revealing a serious potential to obscure traces of real content, including sensitive and harmful materials that could be mistakenly classified as synthetic, thereby undermining the protection of individuals depicted. To address this issue, we propose a two-stage detection pipeline that effectively differentiates between pristine, laundered, and fully synthetic images (those generated from text prompts), showing robustness across various conditions. Finally, we highlight another alarming property of image laundering, which appears to mask the unique artifacts exploited by forensic detectors to solve the camera model identification task, strongly undermining their performance. Our experimental code is available at https://github.com/polimi-ispl/synthetic-image-detection.

The vast accessibility of Synthetic Aperture Radar (SAR) images through online portals has propelled the research across various fields. This widespread use and easy availability have unfortunately made SAR data susceptible to malicious alterations, such as local editing applied to the images for inserting or covering the presence of sensitive targets. Vulnerability is further emphasized by the fact that most SAR products, despite their original complex nature, are often released as amplitude-only information, allowing even inexperienced attackers to edit and easily alter the pixel content. To contrast malicious manipulations, in the last years the forensic community has begun to dig into the SAR manipulation issue, proposing detectors that effectively localize the tampering traces in amplitude images. Nonetheless, in this paper we demonstrate that an expert practitioner can exploit the complex nature of SAR data to obscure any signs of manipulation within a locally altered amplitude image. We refer to this approach as a counter-forensic attack. To achieve the concealment of manipulation traces, the attacker can simulate a re-acquisition of the manipulated scene by the SAR system that initially generated the pristine image. In doing so, the attacker can obscure any evidence of manipulation, making it appear as if the image was legitimately produced by the system. We assess the effectiveness of the proposed counter-forensic approach across diverse scenarios, examining various manipulation operations. The obtained results indicate that our devised attack successfully eliminates traces of manipulation, deceiving even the most advanced forensic detectors.

Due to the latest environmental concerns in keeping at bay contaminants emissions in urban areas, air pollution forecasting has been rising the forefront of all researchers around the world. When predicting pollutant concentrations, it is common to include the effects of environmental factors that influence these concentrations within an extended period, like traffic, meteorological conditions and geographical information. Most of the existing approaches exploit this information as past covariates, i.e., past exogenous variables that affected the pollutant but were not affected by it. In this paper, we present a novel forecasting methodology to predict NO$_2$ concentration via both past and future covariates. Future covariates are represented by weather forecasts and future calendar events, which are already known at prediction time. In particular, we deal with air quality observations in a city-wide network of ground monitoring stations, modeling the data structure and estimating the predictions with a Spatiotemporal Graph Neural Network (STGNN). We propose a conditioning block that embeds past and future covariates into the current observations. After extracting meaningful spatiotemporal representations, these are fused together and projected into the forecasting horizon to generate the final prediction. To the best of our knowledge, it is the first time that future covariates are included in time series predictions in a structured way. Remarkably, we find that conditioning on future weather information has a greater impact than considering past traffic conditions. We release our code implementation at https://github.com/polimi-ispl/MAGCRN., Comment: 5 pages, 4 figures, 1 table, accepted at IEEE-IGARSS 2024

Many deep learning synthetic speech generation tools are readily available. The use of synthetic speech has caused financial fraud, impersonation of people, and misinformation to spread. For this reason forensic methods that can detect synthetic speech have been proposed. Existing methods often overfit on one dataset and their performance reduces substantially in practical scenarios such as detecting synthetic speech shared on social platforms. In this paper we propose, Patched Spectrogram Synthetic Speech Detection Transformer (PS3DT), a synthetic speech detector that converts a time domain speech signal to a mel-spectrogram and processes it in patches using a transformer neural network. We evaluate the detection performance of PS3DT on ASVspoof2019 dataset. Our experiments show that PS3DT performs well on ASVspoof2019 dataset compared to other approaches using spectrogram for synthetic speech detection. We also investigate generalization performance of PS3DT on In-the-Wild dataset. PS3DT generalizes well than several existing methods on detecting synthetic speech from an out-of-distribution dataset. We also evaluate robustness of PS3DT to detect telephone quality synthetic speech and synthetic speech shared on social platforms (compressed speech). PS3DT is robust to compression and can detect telephone quality synthetic speech better than several existing methods., Comment: Accepted as long oral paper at ICMLA 2023

Recent advancements in synthetic speech generation have led to the creation of forged audio data that are almost indistinguishable from real speech. This phenomenon poses a new challenge for the multimedia forensics community, as the misuse of synthetic media can potentially cause adverse consequences. Several methods have been proposed in the literature to mitigate potential risks and detect synthetic speech, mainly focusing on the analysis of the speech itself. However, recent studies have revealed that the most crucial frequency bands for detection lie in the highest ranges (above 6000 Hz), which do not include any speech content. In this work, we extensively explore this aspect and investigate whether synthetic speech detection can be performed by focusing only on the background component of the signal while disregarding its verbal content. Our findings indicate that the speech component is not the predominant factor in performing synthetic speech detection. These insights provide valuable guidance for the development of new synthetic speech detectors and their interpretability, together with some considerations on the existing work in the audio forensics field.

In recent years, Artificial Intelligence (AI) shows a spectacular ability of insertion inside a variety of disciplines which use it for scientific advancements and which sometimes improve it for their conceptual and methodological needs. According to the transverse science framework originally conceived by Shinn and Joerges, AI can be seen as an instrument which is progressively acquiring a universal character through its diffusion across science. In this paper we address empirically one aspect of this diffusion, namely the penetration of AI into a specific field of research. Taking neuroscience as a case study, we conduct a scientometric analysis of the development of AI in this field. We especially study the temporal egocentric citation network around the articles included in this literature, their represented journals and their authors linked together by a temporal collaboration network. We find that AI is driving the constitution of a particular disciplinary ecosystem in neuroscience which is distinct from other subfields, and which is gathering atypical scientific profiles who are coming from neuroscience or outside it. Moreover we observe that this AI community in neuroscience is socially confined in a specific subspace of the neuroscience collaboration network, which also publishes in a small set of dedicated journals that are mostly active in AI research. According to these results, the diffusion of AI in a discipline such as neuroscience didn't really challenge its disciplinary orientations but rather induced the constitution of a dedicated socio-cognitive environment inside this field.