In an era where single large language models have dominated the landscape of artificial intelligence for years, multi-agent systems arise as new protagonists in conversational task-solving. While previous studies have showcased their potential in reasoning tasks and creative endeavors, an analysis of their limitations concerning the conversational paradigms and the impact of individual agents is missing. It remains unascertained how multi-agent discussions perform across tasks of varying complexity and how the structure of these conversations influences the process. To fill that gap, this work systematically evaluates multi-agent systems across various discussion paradigms, assessing their strengths and weaknesses in both generative tasks and question-answering tasks. Alongside the experiments, I propose a taxonomy of 20 multi-agent research studies from 2022 to 2024, followed by the introduction of a framework for deploying multi-agent LLMs in conversational task-solving. I demonstrate that while multi-agent systems excel in complex reasoning tasks, outperforming a single model by leveraging expert personas, they fail on basic tasks. Concretely, I identify three challenges that arise: 1) While longer discussions enhance reasoning, agents fail to maintain conformity to strict task requirements, which leads to problem drift, making shorter conversations more effective for basic tasks. 2) Prolonged discussions risk alignment collapse, raising new safety concerns for these systems. 3) I showcase discussion monopolization through long generations, posing the problem of fairness in decision-making for tasks like summarization. This work uncovers both the potential and challenges that arise with multi-agent interaction and varying conversational paradigms, providing insights into how future research could improve the efficiency, performance, and safety of multi-agent LLMs.

We aim to evaluate Large Language Models (LLMs) for embodied decision making. While a significant body of work has been leveraging LLMs for decision making in embodied environments, we still lack a systematic understanding of their performance because they are usually applied in different domains, for different purposes, and built based on different inputs and outputs. Furthermore, existing evaluations tend to rely solely on a final success rate, making it difficult to pinpoint what ability is missing in LLMs and where the problem lies, which in turn blocks embodied agents from leveraging LLMs effectively and selectively. To address these limitations, we propose a generalized interface (Embodied Agent Interface) that supports the formalization of various types of tasks and input-output specifications of LLM-based modules. Specifically, it allows us to unify 1) a broad set of embodied decision-making tasks involving both state and temporally extended goals, 2) four commonly-used LLM-based modules for decision making: goal interpretation, subgoal decomposition, action sequencing, and transition modeling, and 3) a collection of fine-grained metrics which break down evaluation into various types of errors, such as hallucination errors, affordance errors, various types of planning errors, etc. Overall, our benchmark offers a comprehensive assessment of LLMs' performance for different subtasks, pinpointing the strengths and weaknesses in LLM-powered embodied AI systems, and providing insights for effective and selective use of LLMs in embodied decision making.

We present Multi-expert Prompting, a novel enhancement of ExpertPrompting (Xu et al., 2023), designed to improve the large language model (LLM) generation. Specifically, it guides an LLM to fulfill an input instruction by simulating multiple experts, aggregating their responses, and selecting the best among individual and aggregated responses. This process is performed in a single chain of thoughts through our seven carefully designed subtasks derived from the Nominal Group Technique (Ven and Delbecq, 1974), a well-established decision-making framework. Our evaluations demonstrate that Multi-expert Prompting significantly outperforms ExpertPrompting and comparable baselines in enhancing the truthfulness, factuality, informativeness, and usefulness of responses while reducing toxicity and hurtfulness. It further achieves state-of-the-art truthfulness by outperforming the best baseline by 8.69% with ChatGPT. Multi-expert Prompting is efficient, explainable, and highly adaptable to diverse scenarios, eliminating the need for manual prompt construction.

With text analysis tools becoming increasingly sophisticated over the last decade, researchers now face a decision of whether to use state-of-the-art models that provide high performance but that can be highly opaque in their operations and computationally intensive to run. The alternative, frequently, is to rely on older, manually crafted textual scoring tools that are transparently and easily applied, but can suffer from limited performance. I present an alternative that combines the strengths of both: lexicons created with minimal researcher inputs from generic (pretrained) word embeddings. Presenting a number of conceptual lexicons produced from FastText and GloVe (6B) vector representations of words, I argue that embedding-based lexicons respond to a need for transparent yet high-performance text measuring tools.

Climate change poses critical challenges globally, disproportionately affecting low-income countries that often lack resources and linguistic representation on the international stage. Despite Bangladesh's status as one of the most vulnerable nations to climate impacts, research gaps persist in Bengali-language studies related to climate change and NLP. To address this disparity, we introduce Dhoroni, a novel Bengali (Bangla) climate change and environmental news dataset, comprising a 2300 annotated Bangla news articles, offering multiple perspectives such as political influence, scientific/statistical data, authenticity, stance detection, and stakeholder involvement. Furthermore, we present an in-depth exploratory analysis of Dhoroni and introduce BanglaBERT-Dhoroni family, a novel baseline model family for climate and environmental opinion detection in Bangla, fine-tuned on our dataset. This research contributes significantly to enhancing accessibility and analysis of climate discourse in Bengali (Bangla), addressing crucial communication and research gaps in climate-impacted regions like Bangladesh with 180 million people.

On Monday, Apple released new software that will enrich the lives of iPhone-donning journalists everywhere--and anyone who needs to make sure they get the exact details of a conversation right: a built-in call recording system. The new feature, simply dubbed Call Recording, allows users to record a phone call by clicking a button. From there, the system will alert the recipient that recording has started by saying, "This call will be recorded." This doesn't appear optional--if you record, the other people on the line will be notified. Each year, I pay 89.99--a disgusting amount of money--for a service called TapeACall.

If the presidential election has provided relief from anything, it has been the generative-AI boom. Neither Kamala Harris nor Donald Trump has made much of the technology in their public messaging, and they have not articulated particularly detailed AI platforms. Bots do not seem to rank among the economy, immigration, abortion rights, and other issues that can make or break campaigns. Americans are very invested, and very worried, about the future of artificial intelligence. Polling consistently shows that a majority of adults from both major parties support government regulation of AI, and that demand for regulation might even be growing.

At its best, AI search can better infer a user's intent, amplify quality content, and synthesize information from diverse sources. But if AI search becomes our primary portal to the web, it threatens to disrupt an already precarious digital economy. Today, the production of content online depends on a fragile set of incentives tied to virtual foot traffic: ads, subscriptions, donations, sales, or brand exposure. By shielding the web behind an all-knowing chatbot, AI search could deprive creators of the visits and "eyeballs" they need to survive. If AI search breaks up this ecosystem, existing law is unlikely to help.

AI chatbots impersonating Molly Russell and Brianna Ghey have been found on the controversial site Character.ai. Brianna Ghey was murdered by two teenagers in 2023 while Molly Russell took her own life at the age of 14 after viewing self-harm-related content on social media. In an act described as'sickening', the site's users employed the girl's names, pictures, and biographical details to create dozens of automated bots. Despite violating the site's terms of service, these imitation avatars posing as the two girls were allowed to amass thousands of chats. One impersonating Molly Russell even claimed to be an'expert on the final years of Molly's life'.

We introduce the Autoregressive PDE Emulator Benchmark (APEBench), a comprehensive benchmark suite to evaluate autoregressive neural emulators for solving partial differential equations. APEBench is based on JAX and provides a seamlessly integrated differentiable simulation framework employing efficient pseudo-spectral methods, enabling 46 distinct PDEs across 1D, 2D, and 3D. Facilitating systematic analysis and comparison of learned emulators, we propose a novel taxonomy for unrolled training and introduce a unique identifier for PDE dynamics that directly relates to the stability criteria of classical numerical methods. APEBench enables the evaluation of diverse neural architectures, and unlike existing benchmarks, its tight integration of the solver enables support for differentiable physics training and neural-hybrid emulators. Moreover, APEBench emphasizes rollout metrics to understand temporal generalization, providing insights into the long-term behavior of emulating PDE dynamics. In several experiments, we highlight the similarities between neural emulators and numerical simulators.