This paper presents a systematic review of the literature on evaluation criteria for Trustworthy Artificial Intelligence (TAI), with a focus on the seven EU principles of TAI. This systematic literature review identifies and analyses current evaluation criteria, maps them to the EU TAI principles and proposes a new classification system for each principle. The findings reveal both a need for and significant barriers to standardising criteria for TAI evaluation. The proposed classification contributes to the development, selection and standardization of evaluation criteria for TAI governance.

Automatic refinement of patent claims in patent applications is crucial from the perspective of intellectual property strategy. In this paper, we propose ClaimBrush, a novel framework for automated patent claim refinement that includes a dataset and a rewriting model. We constructed a dataset for training and evaluating patent claim rewriting models by collecting a large number of actual patent claim rewriting cases from the patent examination process. Using the constructed dataset, we built an automatic patent claim rewriting model by fine-tuning a large language model. Furthermore, we enhanced the performance of the automatic patent claim rewriting model by applying preference optimization based on a prediction model of patent examiners' Office Actions. The experimental results showed that our proposed rewriting model outperformed heuristic baselines and zero-shot learning in state-of-the-art large language models. Moreover, preference optimization based on patent examiners' preferences boosted the performance of patent claim refinement.

The objective of this paper is to leverage insights from Alignment Theory (AT) research, which primarily focus on the potential pitfalls of technical alignment in Artificial Intelligence, to critically examine the European Union's Artificial Intelligence Act (EU AI Act). In the context of AT research, several key failure modes - such as proxy gaming, goal drift, reward hacking or specification gaming - have been identified. These can arise when AI systems are not properly aligned with their intended objectives. The central logic of this report is: what can we learn if we treat regulatory efforts in the same way as we treat advanced AI systems? By applying these concepts to the EU AI Act, this project uncovers potential vulnerabilities and areas for improvement in the regulation, ensuring it effectively addresses the complexities and risks associated with AI technologies.

The emerging field of AI-generated art has witnessed the rise of prompt marketplaces, where creators can purchase, sell, or share prompts for generating unique artworks. These marketplaces often assert ownership over prompts, claiming them as intellectual property. This paper investigates whether concealed prompts sold on prompt marketplaces can be considered as secure intellectual property, given that humans and AI tools may be able to approximately infer the prompts based on publicly advertised sample images accompanying each prompt on sale. Specifically, our survey aims to assess (i) how accurately can humans infer the original prompt solely by examining an AI-generated image, with the goal of generating images similar to the original image, and (ii) the possibility of improving upon individual human and AI prompt inferences by crafting human-AI combined prompts with the help of a large language model. Although previous research has explored the use of AI and machine learning to infer (and also protect against) prompt inference, we are the first to include humans in the loop. Our findings indicate that while humans and human-AI collaborations can infer prompts and generate similar images with high accuracy, they are not as successful as using the original prompt.

We develop four versions of GreekLegalRoBERTa, which are four large language models trained on Greek legal and nonlegal text. We show that our models surpass the performance of GreekLegalBERT, Greek- LegalBERT-v2, and GreekBERT in two tasks involving Greek legal documents: named entity recognition and multi-class legal topic classification. We view our work as a contribution to the study of domain-specific NLP tasks in low-resource languages, like Greek, using modern NLP techniques and methodologies.

Smart vehicles produce large amounts of data, much of which is sensitive and at risk of privacy breaches. As attackers increasingly exploit anonymised metadata within these datasets to profile drivers, it's important to find solutions that mitigate this information leakage without hindering innovation and ongoing research. Synthetic data has emerged as a promising tool to address these privacy concerns, as it allows for the replication of real-world data relationships while minimising the risk of revealing sensitive information. In this paper, we examine the use of synthetic data to tackle these challenges. We start by proposing a comprehensive taxonomy of 14 in-vehicle sensors, identifying potential attacks and categorising their vulnerability. We then focus on the most vulnerable signals, using the Passive Vehicular Sensor (PVS) dataset to generate synthetic data with a Tabular Variational Autoencoder (TVAE) model, which included over 1 million data points. Finally, we evaluate this against 3 core metrics: fidelity, utility, and privacy. Our results show that we achieved 90.1% statistical similarity and 78% classification accuracy when tested on its original intent while also preventing the profiling of the driver. The code can be found at https://github.com/krish-parikh/Synthetic-Data-Generation

There have been many benchmarks for evaluating long-context language models (LCLMs), but developers often rely on synthetic tasks like needle-in-a-haystack (NIAH) or arbitrary subsets of tasks. It remains unclear whether they translate to the diverse downstream applications of LCLMs, and the inconsistency further complicates model comparison. We investigate the underlying reasons behind current practices and find that existing benchmarks often provide noisy signals due to low coverage of applications, insufficient lengths, unreliable metrics, and incompatibility with base models. In this work, we present HELMET (How to Evaluate Long-context Models Effectively and Thoroughly), a comprehensive benchmark encompassing seven diverse, application-centric categories. We also address many issues in previous benchmarks by adding controllable lengths up to 128k tokens, model-based evaluation for reliable metrics, and few-shot prompting for robustly evaluating base models. Consequently, we demonstrate that HELMET offers more reliable and consistent rankings of frontier LCLMs. Through a comprehensive study of 51 LCLMs, we find that (1) synthetic tasks like NIAH are not good predictors of downstream performance; (2) the diverse categories in HELMET exhibit distinct trends and low correlation with each other; and (3) while most LCLMs achieve perfect NIAH scores, open-source models significantly lag behind closed ones when the task requires full-context reasoning or following complex instructions -- the gap widens with increased lengths. Finally, we recommend using our RAG tasks for fast model development, as they are easy to run and more predictive of other downstream performance; ultimately, we advocate for a holistic evaluation across diverse tasks.

Text-to-image diffusion models rely on massive, web-scale datasets. Training them from scratch is computationally expensive, and as a result, developers often prefer to make incremental updates to existing models. These updates often compose fine-tuning steps (to learn new concepts or improve model performance) with "unlearning" steps (to "forget" existing concepts, such as copyrighted works or explicit content). In this work, we demonstrate a critical and previously unknown vulnerability that arises in this paradigm: even under benign, non-adversarial conditions, fine-tuning a text-to-image diffusion model on seemingly unrelated images can cause it to "relearn" concepts that were previously "unlearned." We comprehensively investigate the causes and scope of this phenomenon, which we term concept resurgence, by performing a series of experiments which compose "mass concept erasure" (the current state of the art for unlearning in text-to-image diffusion models (Lu et al., 2024)) with subsequent fine-tuning of Stable Diffusion v1.4. Our findings underscore the fragility of composing incremental model updates, and raise serious new concerns about current approaches to ensuring the safety and alignment of text-to-image diffusion models.

The rise of AI-generated content in popular information sources raises significant concerns about accountability, accuracy, and bias amplification. Beyond directly impacting consumers, the widespread presence of this content poses questions for the long-term viability of training language models on vast internet sweeps. We use GPTZero, a proprietary AI detector, and Binoculars, an open-source alternative, to establish lower bounds on the presence of AI-generated content in recently created Wikipedia pages. Both detectors reveal a marked increase in AI-generated content in recent pages compared to those from before the release of GPT-3.5. With thresholds calibrated to achieve a 1% false positive rate on pre-GPT-3.5 articles, detectors flag over 5% of newly created English Wikipedia articles as AI-generated, with lower percentages for German, French, and Italian articles. Flagged Wikipedia articles are typically of lower quality and are often self-promotional or partial towards a specific viewpoint on controversial topics.

Pretrained Deep Neural Networks (DNNs), developed from extensive datasets to integrate multifaceted knowledge, are increasingly recognized as valuable intellectual property (IP). To safeguard these models against IP infringement, strategies for ownership verification and usage authorization have emerged. Unlike most existing IP protection strategies that concentrate on restricting direct access to the model, our study addresses an extended DNN IP issue: applicability authorization, aiming to prevent the misuse of learned knowledge, particularly in unauthorized transfer learning scenarios. We propose Non-Transferable Pruning (NTP), a novel IP protection method that leverages model pruning to control a pretrained DNN's transferability to unauthorized data domains. Selective pruning can deliberately diminish a model's suitability on unauthorized domains, even with full fine-tuning. Specifically, our framework employs the alternating direction method of multipliers (ADMM) for optimizing both the model sparsity and an innovative non-transferable learning loss, augmented with Fisher space discriminative regularization, to constrain the model's generalizability to the target dataset. We also propose a novel effective metric to measure the model non-transferability: Area Under the Sample-wise Learning Curve (SLC-AUC). This metric facilitates consideration of full fine-tuning across various sample sizes. Experimental results demonstrate that NTP significantly surpasses the state-of-the-art non-transferable learning methods, with an average SLC-AUC at $-0.54$ across diverse pairs of source and target domains, indicating that models trained with NTP do not suit for transfer learning to unauthorized target domains. The efficacy of NTP is validated in both supervised and self-supervised learning contexts, confirming its applicability in real-world scenarios.