With the emergence of widely available powerful LLMs, disinformation generated by large Language Models (LLMs) has become a major concern. Historically, LLM detectors have been touted as a solution, but their effectiveness in the real world is still to be proven. In this paper, we focus on an important setting in information operations -- short news-like posts generated by moderately sophisticated attackers. We demonstrate that existing LLM detectors, whether zero-shot or purpose-trained, are not ready for real-world use in that setting. All tested zero-shot detectors perform inconsistently with prior benchmarks and are highly vulnerable to sampling temperature increase, a trivial attack absent from recent benchmarks. A purpose-trained detector generalizing across LLMs and unseen attacks can be developed, but it fails to generalize to new human-written texts. We argue that the former indicates domain-specific benchmarking is needed, while the latter suggests a trade-off between the adversarial evasion resilience and overfitting to the reference human text, with both needing evaluation in benchmarks and currently absent. We believe this suggests a re-consideration of current LLM detector benchmarking approaches and provides a dynamically extensible benchmark to allow it (https://github.com/Reliable-Information-Lab-HEVS/dynamic_llm_detector_benchmark).

Federated Learning (FL) has gained significant popularity due to its effectiveness in training machine learning models across diverse sites without requiring direct data sharing. While various algorithms along with their optimization analyses have shown that FL with local updates is a communication-efficient distributed learning framework, the generalization performance of FL with local updates has received comparatively less attention. This lack of investigation can be attributed to the complex interplay between data heterogeneity and infrequent communication due to the local updates within the FL framework. This motivates us to investigate a fundamental question in FL: Can we quantify the impact of data heterogeneity and local updates on the generalization performance for FL as the learning process evolves? To this end, we conduct a comprehensive theoretical study of FL's generalization performance using a linear model as the first step, where the data heterogeneity is considered for both the stationary and online/non-stationary cases. By providing closed-form expressions of the model error, we rigorously quantify the impact of the number of the local updates (denoted as $K$) under three settings ($K=1$, $K<\infty$, and $K=\infty$) and show how the generalization performance evolves with the number of rounds $t$. Our investigation also provides a comprehensive understanding of how different configurations (including the number of model parameters $p$ and the number of training samples $n$) contribute to the overall generalization performance, thus shedding new insights (such as benign overfitting) for implementing FL over networks.

In many applications of multiple testing, ancillary information is available, reflecting the hypothesis null or alternative status. Several methods have been developed to leverage this ancillary information to enhance testing power, typically requiring the ancillary information is helpful enough to ensure favorable performance. In this paper, we develop a robust and effective distance-assisted multiple testing procedure named DART2, designed to be powerful and robust regardless of the quality of ancillary information. When the ancillary information is helpful, DART2 can asymptotically control FDR while improving power; otherwise, DART2 can still control FDR and maintain power at least as high as ignoring the ancillary information. We demonstrated DART2's superior performance compared to existing methods through numerical studies under various settings. In addition, DART2 has been applied to a gene association study where we have shown its superior accuracy and robustness under two different types of ancillary information.

Large language models (LLMs), while memorizing parts of their training data scraped from the Internet, may also inadvertently encode societal preferences and norms. As these models are integrated into sociotechnical systems, it is crucial that the norms they encode align with societal expectations. These norms could vary across models, hyperparameters, optimization techniques, and datasets. This is especially challenging due to prompt sensitivity$-$small variations in prompts yield different responses, rendering existing assessment methodologies unreliable. There is a need for a comprehensive framework covering various models, optimization, and datasets, along with a reliable methodology to assess encoded norms. We present LLM-CI, the first open-sourced framework to assess privacy norms encoded in LLMs. LLM-CI uses a Contextual Integrity-based factorial vignette methodology to assess the encoded norms across different contexts and LLMs. We propose the multi-prompt assessment methodology to address prompt sensitivity by assessing the norms from only the prompts that yield consistent responses across multiple variants. Using LLM-CI and our proposed methodology, we comprehensively evaluate LLMs using IoT and COPPA vignettes datasets from prior work, examining the impact of model properties (e.g., hyperparameters, capacity) and optimization strategies (e.g., alignment, quantization).

Detecting false information on social media is critical in mitigating its negative societal impacts. To reduce the propagation of false information, automated detection provide scalable, unbiased, and cost-effective methods. However, there are three potential research areas identified which once solved improve detection. First, current AI-based solutions often provide a uni-dimensional analysis on a complex, multi-dimensional issue, with solutions differing based on the features used. Furthermore, these methods do not account for the temporal and dynamic changes observed within the document's life cycle. Second, there has been little research on the detection of coordinated information campaigns and in understanding the intent of the actors and the campaign. Thirdly, there is a lack of consideration of cross-platform analysis, with existing datasets focusing on a single platform, such as X, and detection models designed for specific platform. This work aims to develop methods for effective detection of false information and its propagation. To this end, firstly we aim to propose the creation of an ensemble multi-faceted framework that leverages multiple aspects of false information. Secondly, we propose a method to identify actors and their intent when working in coordination to manipulate a narrative. Thirdly, we aim to analyse the impact of cross-platform interactions on the propagation of false information via the creation of a new dataset.

This paper presents the experiments and results for the CheckThat! Lab at CLEF 2024 Task 6: Robustness of Credibility Assessment with Adversarial Examples (InCrediblAE). The primary objective of this task was to generate adversarial examples in five problem domains in order to evaluate the robustness of widely used text classification methods (fine-tuned BERT, BiLSTM, and RoBERTa) when applied to credibility assessment issues. This study explores the application of ensemble learning to enhance adversarial attacks on natural language processing (NLP) models. We systematically tested and refined several adversarial attack methods, including BERT-Attack, Genetic algorithms, TextFooler, and CLARE, on five datasets across various misinformation tasks. By developing modified versions of BERT-Attack and hybrid methods, we achieved significant improvements in attack effectiveness. Our results demonstrate the potential of modification and combining multiple methods to create more sophisticated and effective adversarial attack strategies, contributing to the development of more robust and secure systems.

Predicting the performance of LLMs on individual task instances is essential to ensure their reliability in high-stakes applications. To do so, a possibility is to evaluate the considered LLM on a set of task instances and train an assessor to predict its performance based on features of the instances. However, this approach requires evaluating each new LLM on a sufficiently large set of task instances to train an assessor specific to it. In this work, we leverage the evaluation results of previously tested LLMs to reduce the number of evaluations required to predict the performance of a new LLM. In practice, we propose to test the new LLM on a small set of reference instances and train a generic assessor which predicts the performance of the LLM on an instance based on the performance of the former on the reference set and features of the instance of interest. We conduct empirical studies on HELM-Lite and KindsOfReasoning, a collection of existing reasoning datasets that we introduce, where we evaluate all instruction-fine-tuned OpenAI models until the January 2024 version of GPT4. When predicting performance on instances with the same distribution as those used to train the generic assessor, we find this achieves performance comparable to the LLM-specific assessors trained on the full set of instances. Additionally, we find that randomly selecting the reference instances performs as well as some advanced selection methods we tested. For out of distribution, however, no clear winner emerges and the overall performance is worse, suggesting that the inherent predictability of LLMs is low.

Isidore of Seville is credited with the adage that it is language that gives birth to a people, and not the other way around , underlining the profound role played by language in the formation of cultural and social identity. Today, of the more than 7100 languages listed, a significant number are endangered. Since the 1970s, linguists, information seekers and enthusiasts have helped develop digital resources and automatic tools to support a wide range of languages, including endangered ones. The advent of Large Language Model (LLM) technologies holds both promise and peril. They offer unprecedented possibilities for the translation and generation of content and resources, key elements in the preservation and revitalisation of languages. They also present threat of homogenisation, cultural oversimplification and the further marginalisation of already vulnerable languages. The talk this paper is based on has proposed an initiatory journey, exploring the potential paths and partnerships between technology and tradition, with a particular focus on the Occitan language. Occitan is a language from Southern France, parts of Spain and Italy that played a major cultural and economic role, particularly in the Middle Ages. It is now endangered according to UNESCO. The talk critically has examined how human expertise and artificial intelligence can work together to offer hope for preserving the linguistic diversity that forms the foundation of our global and especially our European heritage while addressing some of the ethical and practical challenges that accompany the use of these powerful technologies. This paper is based on the keynote I gave at the 46th European Conference on Information Retrieval (ECIR 2024). As an alternative to reading this paper, a video talk is available online. 1 Date: 26 March 2024.

OutKick host Tomi Lahren sounds off on viral video of Amazon Alexa praising Harris while refusing to speak on why its users should support Trump. While videos circulate online of Amazon's Alexa giving vastly different answers when asked to make a quick argument for voting for Vice President Harris versus voting for former President Trump, federal donor records show the tech giant's employees have a clear favorite. A review of a Federal Election Commission database shows employees at Jeff Bezos' companies Amazon and Blue Origin have contributed significantly to VP Kamala Harris' campaign. According to the nonpartisan research group OpenSecrets, Amazon donors have contributed 1,000,140 to Vice President Kamala Harris during the 2024 election cycle. Blue Origin workers donated much less to the Harris' campaign, at roughly 27,000.

OutKick host Tomi Lahren sounds off on viral video of Amazon Alexa praising Harris while refusing to speak on why its users should support Trump. FIRST ON FOX: Sen. Lindsey Graham, R-S.C., issued a stern warning to Amazon after its virtual assistant technology, Alexa, was found to be politically biased in favor of Vice President Kamala Harris over former President Trump. Graham, the ranking member of the Senate Committee on the Judiciary, told Amazon president and CEO Andrew Jassy in a letter he was putting him "on notice that I will not allow this to go unaddressed." On Tuesday, videos of interactions with Alexa went viral as the technology responded to queries on why someone should vote for Harris or why they should vote for Trump. Sen. Lindsey Graham, R-S.C., demanded an explanation from Amazon after its Alexa technology gave biased answers.