As Vision and Language models (VLMs) become accessible across the globe, it is important that they demonstrate cultural knowledge. In this paper, we introduce CROPE, a visual question answering benchmark designed to probe the knowledge of culture-specific concepts and evaluate the capacity for cultural adaptation through contextual information. This allows us to distinguish between parametric knowledge acquired during training and contextual knowledge provided during inference via visual and textual descriptions. Our evaluation of several state-of-the-art open VLMs shows large performance disparities between culture-specific and common concepts in the parametric setting. Moreover, experiments with contextual knowledge indicate that models struggle to effectively utilize multimodal information and bind culture-specific concepts to their depictions. Our findings reveal limitations in the cultural understanding and adaptability of current VLMs that need to be addressed toward more culturally inclusive models.

In what may be a first, the US Department of Justice this week charged a hacker with attempting to cause injury and death by launching distributed denial-of-service (DDoS) attacks against hospitals. Ahmed Omer and his brother Alaa are accused of carrying out a cyberattack spree that targeted hundreds of victims under the hacktivist banner Anonymous Sudan. The group's DDoS victims included Microsoft's Azure cloud services, OpenAI's ChatGPT, and Israel's missile alert system, according to prosecutors. It was the brothers' alleged attacks on hospitals, however, that drew the most serious accusations from the Justice Department, which singled out Ahmed for allegedly seeking to kill people with the crude cyberattacks that overwhelm systems, knocking them offline. If someone told you there's a tool that can--using only open source information--create a "cyber profile" of you that can locate your phone in real time or place you at the scene of a crime at any date in the past, would you believe them?

With the recent emergence of foundation models trained on internet-scale data and demonstrating remarkable generalization capabilities, such foundation models have become more widely adopted, leading to an expanding range of application domains. Despite this rapid proliferation, the trustworthiness of foundation models remains underexplored. Specifically, the out-of-distribution detection (OoDD) capabilities of large vision-language models (LVLMs), such as GPT-4o, which are trained on massive multi-modal data, have not been sufficiently addressed. The disparity between their demonstrated potential and practical reliability raises concerns regarding the safe and trustworthy deployment of foundation models. To address this gap, we evaluate and analyze the OoDD capabilities of various proprietary and open-source LVLMs. Our investigation contributes to a better understanding of how these foundation models represent confidence scores through their generated natural language responses. Based on our observations, we propose a self-guided prompting approach, termed \emph{Reflexive Guidance (ReGuide)}, aimed at enhancing the OoDD capability of LVLMs by leveraging self-generated image-adaptive concept suggestions. Experimental results demonstrate that our ReGuide enhances the performance of current LVLMs in both image classification and OoDD tasks.

As the use of large language models (LLMs) for code generation becomes more prevalent in software development, it is critical to enhance both the efficiency and correctness of the generated code. Existing methods and models primarily focus on the correctness of LLM-generated code, ignoring efficiency. In this work, we present Effi-Code, an approach to enhancing code generation in LLMs that can improve both efficiency and correctness. We introduce a Self-Optimization process based on Overhead Profiling that leverages open-source LLMs to generate a high-quality dataset of correct and efficient code samples. This dataset is then used to fine-tune various LLMs. Our method involves the iterative refinement of generated code, guided by runtime performance metrics and correctness checks. Extensive experiments demonstrate that models fine-tuned on the Effi-Code show significant improvements in both code correctness and efficiency across task types. For example, the pass@1 of DeepSeek-Coder-6.7B-Instruct generated code increases from \textbf{43.3\%} to \textbf{76.8\%}, and the average execution time for the same correct tasks decreases by \textbf{30.5\%}. Effi-Code offers a scalable and generalizable approach to improving code generation in AI systems, with potential applications in software development, algorithm design, and computational problem-solving. The source code of Effi-Code was released in \url{https://github.com/huangd1999/Effi-Code}.

To model Callon's position in the form of an LLM, one need only train the model on the already available textual corpus of post-industrial Western capitalism, given that any performed instance of the discourse token "economics" or "the economy" will form the statistically average center of attention for an associative network of other, token-level terms. However, although token-level linguistic performatives of the kind described by Callon have played a historically outsized role in the economies of capitalist states, the power of the performative token does not exhaust the concept of economics as a field of human "social creativity", understood as the linguistically/symbolically mediated, historically/mythologically self-consciousness agency of intelligent beings conceptualizing and transforming their own conditions of existence (Graeber 2012). Marcel Mauss, on the other hand, acknowledged the formal autonomy of generative exchange as an existentially human practice that took up various "forms and reasons" across different cases, opening the possibility for theorizing type-level conceptual distinctions based on their functional parallelism across diverse societies, and, in Mauss's own radical argument, even identifying deficiencies in the dominant form of exchange from the perspective of non-dominant forms. Insofar as reflective attention to ethnographic type-tokens like kula, potlach, or mana enhances human economic anthropologists' capacity to recognize patterns of value-creation and transformation within any new set of ethnographic data, a Maussian methodology can meaningfully inform the mechanics of machine learning and provide a touchstone for the integration of anthropological knowledge with AI research. In a future publication, Sheldon will elaborate on this contrast between the "flat ontology" of Actor Network Theory and the "depth ontology" that continues to be generatively employed by logicians, mathematicians, and computer scientists (as well as mystics, magicians, and illusionists), both ancient and modern.

Modern large language models (LLMs) have exhibited cooperative synergy on complex task-solving, and collective decision-making (CDM) is a pivotal component in LLM-based multi-agent collaboration frameworks. Our survey on 52 recent such systems uncovers a severe lack of diversity, with a heavy reliance on dictatorial and plurality voting for CDM. Through the lens of social choice theory, we scrutinize widely-adopted CDM methods and identify their limitations. To enrich current landscape of LLM-based CDM, we present GEDI, an electoral CDM module that incorporates various ordinal preferential voting mechanisms. Our empirical case study across three benchmarks shows that the integration of certain CDM methods can markedly improve the reasoning capabilities and robustness of some leading LLMs, all without requiring intricate system designs. Additionally, we find that some CDM mechanisms generate positive synergies even with as few as three agents. The voting-based methods also demonstrate robustness against single points of failure, as well as diversity in terms of hit-rate@k and subject-wise impacts.

As Large Language Models (LLMs) become increasingly integrated into various facets of society, a significant portion of online text consequently become synthetic. This raises concerns about bias amplification, a phenomenon where models trained on synthetic data amplify the pre-existing biases over successive training iterations. Previous literature seldom discusses bias amplification as an independent issue from model collapse. In this work, we address the gap in understanding the bias amplification of LLMs with four main contributions. Firstly, we propose a theoretical framework, defining the necessary and sufficient conditions for its occurrence, and emphasizing that it occurs independently of model collapse. Using statistical simulations with weighted maximum likelihood estimation, we demonstrate the framework and show how bias amplification arises without the sampling and functional form issues that typically drive model collapse. Secondly, we conduct experiments with GPT-2 to empirically demonstrate bias amplification, specifically examining open-ended generational political bias with a benchmark we developed. We observe that GPT-2 exhibits a right-leaning bias in sentence continuation tasks and that the bias progressively increases with iterative fine-tuning on synthetic data generated by previous iterations. Thirdly, we explore three potential mitigation strategies: Overfitting, Preservation, and Accumulation. We find that both Preservation and Accumulation effectively mitigate bias amplification and model collapse. Finally, using novel mechanistic interpretation techniques, we demonstrate that in the GPT-2 experiments, bias amplification and model collapse are driven by distinct sets of neurons, which aligns with our theoretical framework.

The proliferation of online micro-video platforms has underscored the necessity for advanced recommender systems to mitigate information overload and deliver tailored content. Despite advancements, accurately and promptly capturing dynamic user interests remains a formidable challenge. Inspired by the Platonic Representation Hypothesis, which posits that different data modalities converge towards a shared statistical model of reality, we introduce DreamUMM (Dreaming User Multi-Modal Representation), a novel approach leveraging user historical behaviors to create real-time user representation in a multimoda space. DreamUMM employs a closed-form solution correlating user video preferences with multimodal similarity, hypothesizing that user interests can be effectively represented in a unified multimodal space. Additionally, we propose Candidate-DreamUMM for scenarios lacking recent user behavior data, inferring interests from candidate videos alone. Extensive online A/B tests demonstrate significant improvements in user engagement metrics, including active days and play count. The successful deployment of DreamUMM in two micro-video platforms with hundreds of millions of daily active users, illustrates its practical efficacy and scalability in personalized micro-video content delivery. Our work contributes to the ongoing exploration of representational convergence by providing empirical evidence supporting the potential for user interest representations to reside in a multimodal space.

Neural Algorithmic Reasoning (NAR) research has demonstrated that graph neural networks (GNNs) could learn to execute classical algorithms. However, most previous approaches have always used a recurrent architecture, where each iteration of the GNN matches an iteration of the algorithm. In this paper we study neurally solving algorithms from a different perspective: since the algorithm's solution is often an equilibrium, it is possible to find the solution directly by solving an equilibrium equation. Our approach requires no information on the ground-truth number of steps of the algorithm, both during train and test time. Furthermore, the proposed method improves the performance of GNNs on executing algorithms and is a step towards speeding up existing NAR models. Our empirical evidence, leveraging algorithms from the CLRS-30 benchmark, validates that one can train a network to solve algorithmic problems by directly finding the equilibrium. We discuss the practical implementation of such models and propose regularisations to improve the performance of these equilibrium reasoners.

Retrieval Augmented Language Models (RALMs) have gained significant attention for their ability to generate accurate answer and improve efficiency. However, RALMs are inherently vulnerable to imperfect information due to their reliance on the imperfect retriever or knowledge source. We identify three common scenarios-unanswerable, adversarial, conflicting-where retrieved document sets can confuse RALM with plausible real-world examples. We present the first comprehensive investigation to assess how well RALMs detect and handle such problematic scenarios. Among these scenarios, to systematically examine adversarial robustness we propose a new adversarial attack method, Generative model-based ADVersarial attack (GenADV) and a novel metric Robustness under Additional Document (RAD). Our findings reveal that RALMs often fail to identify the unanswerability or contradiction of a document set, which frequently leads to hallucinations. Moreover, we show the addition of an adversary significantly degrades RALM's performance, with the model becoming even more vulnerable when the two scenarios overlap (adversarial+unanswerable). Our research identifies critical areas for assessing and enhancing the robustness of RALMs, laying the foundation for the development of more robust models.