Large Language Models (LLMs) have excelled at language understanding and generating human-level text. However, even with supervised training and human alignment, these LLMs are susceptible to adversarial attacks where malicious users can prompt the model to generate undesirable text. LLMs also inherently encode potential biases that can cause various harmful effects during interactions. Bias evaluation metrics lack standards as well as consensus and existing methods often rely on human-generated templates and annotations which are expensive and labor intensive. In this work, we train models to automatically create adversarial prompts to elicit biased responses from target LLMs. We present LLM- based bias evaluation metrics and also analyze several existing automatic evaluation methods and metrics. We analyze the various nuances of model responses, identify the strengths and weaknesses of model families, and assess where evaluation methods fall short. We compare these metrics to human evaluation and validate that the LLM-as-a-Judge metric aligns with human judgement on bias in response generation.

Providing recommendations that are both relevant and diverse is a key consideration of modern recommender systems. Optimizing both of these measures presents a fundamental trade-off, as higher diversity typically comes at the cost of relevance, resulting in lower user engagement. Existing recommendation algorithms try to resolve this trade-off by combining the two measures, relevance and diversity, into one aim and then seeking recommendations that optimize the combined objective, for a given number of items to recommend. Traditional approaches, however, do not consider the user interaction with the recommended items. In this paper, we put the user at the central stage, and build on the interplay between relevance, diversity, and user behavior. In contrast to applications where the goal is solely to maximize engagement, we focus on scenarios aiming at maximizing the total amount of knowledge encountered by the user. We use diversity as a surrogate of the amount of knowledge obtained by the user while interacting with the system, and we seek to maximize diversity. We propose a probabilistic user-behavior model in which users keep interacting with the recommender system as long as they receive relevant recommendations, but they may stop if the relevance of the recommended items drops. Thus, for a recommender system to achieve a high-diversity measure, it will need to produce recommendations that are both relevant and diverse. Finally, we propose a novel recommendation strategy that combines relevance and diversity by a copula function. We conduct an extensive evaluation of the proposed methodology over multiple datasets, and we show that our strategy outperforms several state-of-the-art competitors. Our implementation is publicly available at https://github.com/EricaCoppolillo/EXPLORE.

Graph Neural Networks (GNNs) are important across different domains, such as social network analysis and recommendation systems, due to their ability to model complex relational data. This paper introduces subgraph queries as a new task for deep graph learning. Unlike traditional graph prediction tasks that focus on individual components like link prediction or node classification, subgraph queries jointly predict the components of a target subgraph based on evidence that is represented by an observed subgraph. For instance, a subgraph query can predict a set of target links and/or node labels. To answer subgraph queries, we utilize a probabilistic deep Graph Generative Model. Specifically, we inductively train a Variational Graph Auto-Encoder (VGAE) model, augmented to represent a joint distribution over links, node features and labels. Bayesian optimization is used to tune a weighting for the relative importance of links, node features and labels in a specific domain. We describe a deterministic and a sampling-based inference method for estimating subgraph probabilities from the VGAE generative graph distribution, without retraining, in zero-shot fashion. For evaluation, we apply the inference methods on a range of subgraph queries on six benchmark datasets. We find that inference from a model achieves superior predictive performance, surpassing independent prediction baselines with improvements in AUC scores ranging from 0.06 to 0.2 points, depending on the dataset.

We outline the principles of classical assurance for computer-based systems that pose significant risks. We then consider application of these principles to systems that employ Artificial Intelligence (AI) and Machine Learning (ML). A key element in this "dependability" perspective is a requirement to have near-complete understanding of the behavior of critical components, and this is considered infeasible for AI and ML. Hence the dependability perspective aims to minimize trust in AI and ML elements by using "defense in depth" with a hierarchy of less complex systems, some of which may be highly assured conventionally engineered components, to "guard" them. This may be contrasted with the "trustworthy" perspective that seeks to apply assurance to the AI and ML elements themselves. In cyber-physical and many other systems, it is difficult to provide guards that do not depend on AI and ML to perceive their environment (e.g., other vehicles sharing the road with a self-driving car), so both perspectives are needed and there is a continuum or spectrum between them. We focus on architectures toward the dependability end of the continuum and invite others to consider additional points along the spectrum. For guards that require perception using AI and ML, we examine ways to minimize the trust placed in these elements; they include diversity, defense in depth, explanations, and micro-ODDs. We also examine methods to enforce acceptable behavior, given a model of the world. These include classical cyber-physical calculations and envelopes, and normative rules based on overarching principles, constitutions, ethics, or reputation. We apply our perspective to autonomous systems, AI systems for specific functions, generic AI such as Large Language Models, and to Artificial General Intelligence (AGI), and we propose current best practice and an agenda for research.

This is the author's version of the article that has been published in the proceedings of IEEE Visualization conference. The final version of this record is available at: xx.xxxx/TVCG.201x.xxxxxxx/ This metric also provides no visibility into how particular Reconstruction of 3D scenes from 2D images is a technical challenge images, lighting conditions, camera positions, or details of the that impacts domains from Earth and planetary sciences and morphology of the remote environment might interact to create inaccuracies space exploration to augmented and virtual reality. The impact of these unknowns algorithms first identify common features across images compounds in domains where high accuracy terrain reconstruction and then minimize reconstruction errors after estimating the is critical to outcomes, like science or space exploration where there shape of the terrain. This bundle adjustment (BA) step optimizes is no ground truth and inaccurate reconstruction can lead to false around a single, simplifying scalar value that obfuscates many possible results or risking billion-dollar spacecraft.

Recent advancements in long-context modeling have enhanced language models (LMs) for complex tasks across multiple NLP applications. Despite this progress, we find that these models struggle with multi-hop reasoning and exhibit decreased performance in the presence of noisy contexts. In this paper, we introduce Reasoning with Attributions, a novel approach that prompts LMs to supply attributions for each assertion during their reasoning. We validate our approach through experiments on three multi-hop datasets, employing both proprietary and open-source models, and demonstrate its efficacy and resilience. Furthermore, we explore methods to augment reasoning capabilities via fine-tuning and offer an attribution-annotated dataset and a specialized training strategy. Our fine-tuned model achieves competitive performance on multi-hop reasoning benchmarks, closely paralleling proprietary LMs such as ChatGPT and Claude-instant.

For simple classification tasks, we show that users can benefit from the advantages of using small, local, generative language models instead of large commercial models without a trade-off in performance or introducing extra labelling costs. These advantages, including those around privacy, availability, cost, and explainability, are important both in commercial applications and in the broader democratisation of AI. Through experiments on 17 sentence classification tasks (2-4 classes), we show that penalised logistic regression on the embeddings from a small LLM equals (and usually betters) the performance of a large LLM in the "tens-of-shot" regime. This requires no more labelled instances than are needed to validate the performance of the large LLM. Finally, we extract stable and sensible explanations for classification decisions.

On Monday, 404 Media's Samantha Cole reported that the 2.4 trillion company asked workers to download videos from YouTube, Netflix and other datasets to develop commercial AI projects. The graphics card maker is among the tech companies appearing to have adopted a "move fast and break things" ethos as they race to establish dominance in this feverish, too-often-shameful AI gold rush. The training was reportedly to develop models for products like its Omniverse 3D world generator, self-driving car systems and "digital human" efforts. NVIDIA defended its practice in an email to Engadget. The company equated the practice to a person's right to "learn facts, ideas, data, or information from another source and use it to make their own expression."

Artificial intelligence is increasingly being shoved into everything, but AI chatbots in particular have exploded in popularity lately. Companies like OpenAI and Microsoft are pushing the likes of ChatGPT and Copilot to provide you with answers to all of your questions, while companies like Amazon use AI chatbots to guide you through the shopping experience in hopes of scoring sales. But how are people actually using AI chatbots? A new report by The Washington Post aims to answer that question. The newspaper analyzed research data from nearly 200,000 English-language conversations from the WildChat Dataset, a database of over 1 million real-world user conversations with ChatGPT.

Former President Trump told FOX Business' Maria Bartiromo last week that Meta CEO Mark Zuckerberg called him to apologize after Facebook wrongly mislabeled a now-viral photo of the former president. The photo showing Trump raising a fist after a July 13 assassination attempt at his campaign rally in Butler, Pennsylvania, sliced his ear was initially labeled as misinformation on the social media site. "So, Mark Zuckerberg called me. First of all, he called me two times. He called me after the event and he said that was really amazing," Trump told Bartiromo in a "Mornings with Maria" interview that aired Thursday.