In this paper, we propose an autonomous information seeking visual question answering framework, AVIS. Our method leverages a Large Language Model (LLM) to dynamically strategize the utilization of external tools and to investigate their outputs, thereby acquiring the indispensable knowledge needed to provide answers to the posed questions. Responding to visual questions that necessitate external knowledge, such as "What event is commemorated by the building depicted in this image?", is a complex task. This task presents a combinatorial search space that demands a sequence of actions, including invoking APIs, analyzing their responses, and making informed decisions. We conduct a user study to collect a variety of instances of human decision-making when faced with this task. This data is then used to design a system comprised of three components: an LLM-powered planner that dynamically determines which tool to use next, an LLM-powered reasoner that analyzes and extracts key information from the tool outputs, and a working memory component that retains the acquired information throughout the process. The collected user behavior serves as a guide for our system in two key ways. First, we create a transition graph by analyzing the sequence of decisions made by users. This graph delineates distinct states and confines the set of actions available at each state. Second, we use examples of user decision-making to provide our LLM-powered planner and reasoner with relevant contextual instances, enhancing their capacity to make informed decisions. We show that AVIS achieves state-of-the-art results on knowledge-intensive visual question answering benchmarks such as Infoseek and OK-VQA.

Producing high-quality forecasts of key climate variables such as temperature and precipitation on subseasonal time scales has long been a gap in operational forecasting. Recent studies have shown promising results using machine learning (ML) models to advance subseasonal forecasting (SSF), but several open questions remain. First, several past approaches use the average of an ensemble of physics-based forecasts as an input feature of these models. However, ensemble forecasts contain information that can aid prediction beyond only the ensemble mean. Second, past methods have focused on average performance, whereas forecasts of extreme events are far more important for planning and mitigation purposes. Third, climate forecasts correspond to a spatially-varying collection of forecasts, and different methods account for spatial variability in the response differently. Trade-offs between different approaches may be mitigated with model stacking. This paper describes the application of a variety of ML methods used to predict monthly average precipitation and two meter temperature using physics-based predictions (ensemble forecasts) and observational data such as relative humidity, pressure at sea level, or geopotential height, two weeks in advance for the whole continental United States. Regression, quantile regression, and tercile classification tasks using linear models, random forests, convolutional neural networks, and stacked models are considered. The proposed models outperform common baselines such as historical averages (or quantiles) and ensemble averages (or quantiles). This paper further includes an investigation of feature importance, trade-offs between using the full ensemble or only the ensemble average, and different modes of accounting for spatial variability.

The Strait of Gibraltar is a region characterized by intricate oceanic sub-mesoscale features, influenced by topography, tidal forces, instabilities, and nonlinear hydraulic processes, all governed by the nonlinear equations of fluid motion. In this study, we aim to uncover the underlying physics of these phenomena within 3D MIT general circulation model simulations, including waves, eddies, and gyres. To achieve this, we employ Dynamic Mode Decomposition (DMD) to break down simulation snapshots into Koopman modes, with distinct exponential growth/decay rates and oscillation frequencies. Our objectives encompass evaluating DMD's efficacy in capturing known features, unveiling new elements, ranking modes, and exploring order reduction. We also introduce modifications to enhance DMD's robustness, numerical accuracy, and robustness of eigenvalues. DMD analysis yields a comprehensive understanding of flow patterns, internal wave formation, and the dynamics of the Strait of Gibraltar, its meandering behaviors, and the formation of a secondary gyre, notably the Western Alboran Gyre, as well as the propagation of Kelvin and coastal-trapped waves along the African coast. In doing so, it significantly advances our comprehension of intricate oceanographic phenomena and underscores the immense utility of DMD as an analytical tool for such complex datasets, suggesting that DMD could serve as a valuable addition to the toolkit of oceanographers.

FOX News' Greg Palkot reports the latest from the Israel-Lebanon border. A drone attack on a U.S. base in Syria was thwarted on Wednesday, according to a report. Two drones targeting Syria's al-Tanf region were disabled or destroyed by the base defense system, an Iraqi government source told Reuters. The thwarted attack comes as U.S. and Coalition Forces at Combined Joint Task Force Operation Inherent Resolve (CJTF-OIR) installations in Iraq and Syria have been attacked at least 27 times between Oct. 17-31. Of these attacks, 16 happened in Iraq and 11 took place in Syria.

Language models (LMs) often exhibit undesirable text generation behaviors, including generating false, toxic, or irrelevant outputs. Reinforcement learning from human feedback (RLHF) - where human preference judgments on LM outputs are transformed into a learning signal - has recently shown promise in addressing these issues. However, such holistic feedback conveys limited information on long text outputs; it does not indicate which aspects of the outputs influenced user preference; e.g., which parts contain what type(s) of errors. In this paper, we use fine-grained human feedback (e.g., which sentence is false, which sub-sentence is irrelevant) as an explicit training signal. We introduce Fine-Grained RLHF, a framework that enables training and learning from reward functions that are fine-grained in two respects: (1) density, providing a reward after every segment (e.g., a sentence) is generated; and (2) incorporating multiple reward models associated with different feedback types (e.g., factual incorrectness, irrelevance, and information incompleteness). We conduct experiments on detoxification and long-form question answering to illustrate how learning with such reward functions leads to improved performance, supported by both automatic and human evaluation. Additionally, we show that LM behaviors can be customized using different combinations of fine-grained reward models. We release all data, collected human feedback, and codes at https://FineGrainedRLHF.github.io.

Russia said Ukrainian drones damaged a nuclear waste storage facility at the Kursk Nuclear Power Plant on Thursday evening. This comes after the press service for the plant told journalists on Friday that there had been no significant damage from the attacks and that operations were continuing as normal. Intense fighting continued close to the city of Avdiivka, in the Donetsk region of eastern Ukraine. Russia's Ministry of Defence said its air defence systems destroyed 36 Ukraine-launched drones over the Black Sea off the Crimean Peninsula overnight. A statement from the ministry on Telegram did not provide much additional detail.

As cyclones become more intense due to climate change, the rise of AI-based modelling provides a more affordable and accessible approach compared to traditional methods based on mathematical models. This work leverages diffusion models to forecast cyclone trajectories and precipitation patterns by integrating satellite imaging, remote sensing, and atmospheric data, employing a cascaded approach that incorporates forecasting, super-resolution, and precipitation modelling, with training on a dataset of 51 cyclones from six major basins. Experiments demonstrate that the final forecasts from the cascaded models show accurate predictions up to a 36-hour rollout, with SSIM and PSNR values exceeding 0.5 and 20 dB, respectively, for all three tasks. This work also highlights the promising efficiency of AI methods such as diffusion models for high-performance needs, such as cyclone forecasting, while remaining computationally affordable, making them ideal for highly vulnerable regions with critical forecasting needs and financial limitations.

A long-standing goal of AI systems is to perform complex multimodal reasoning like humans. Recently, large language models (LLMs) have made remarkable strides in such multi-step reasoning on the language modality solely by leveraging the chain of thought (CoT) to mimic human thinking. However, the transfer of these advancements to multimodal contexts introduces heightened challenges, including but not limited to the impractical need for labor-intensive annotation and the limitations in terms of flexibility, generalizability, and explainability. To evoke CoT reasoning in multimodality, this work first conducts an in-depth analysis of these challenges posed by multimodality and presents two key insights: "keeping critical thinking" and "letting everyone do their jobs" in multimodal CoT reasoning. Furthermore, this study proposes a novel DDCoT prompting that maintains a critical attitude through negative-space prompting and incorporates multimodality into reasoning by first dividing the reasoning responsibility of LLMs into reasoning and recognition and then integrating the visual recognition capability of visual models into the joint reasoning process. The rationales generated by DDCoT not only improve the reasoning abilities of both large and small language models in zero-shot prompting and fine-tuning learning, significantly outperforming state-of-the-art methods but also exhibit impressive generalizability and explainability.

Languages are known to describe the world in diverse ways. Across lexicons, diversity is pervasive, appearing through phenomena such as lexical gaps and untranslatability. However, in computational resources, such as multilingual lexical databases, diversity is hardly ever represented. In this paper, we introduce a method to enrich computational lexicons with content relating to linguistic diversity. The method is verified through two large-scale case studies on kinship terminology, a domain known to be diverse across languages and cultures: one case study deals with seven Arabic dialects, while the other one with three Indonesian languages. Our results, made available as browseable and downloadable computational resources, extend prior linguistics research on kinship terminology, and provide insight into the extent of diversity even within linguistically and culturally close communities.

Language Models (LMs) often must integrate facts they memorized in pretraining with new information that appears in a given context. These two sources can disagree, causing competition within the model, and it is unclear how an LM will resolve the conflict. On a dataset that queries for knowledge of world capitals, we investigate both distributional and mechanistic determinants of LM behavior in such situations. Specifically, we measure the proportion of the time an LM will use a counterfactual prefix (e.g., "The capital of Poland is London") to overwrite what it learned in pretraining ("Warsaw"). On Pythia and GPT2, the training frequency of both the query country ("Poland") and the in-context city ("London") highly affect the models' likelihood of using the counterfactual. We then use head attribution to identify individual attention heads that either promote the memorized answer or the in-context answer in the logits. By scaling up or down the value vector of these heads, we can control the likelihood of using the in-context answer on new data. This method can increase the rate of generating the in-context answer to 88\% of the time simply by scaling a single head at runtime. Our work contributes to a body of evidence showing that we can often localize model behaviors to specific components and provides a proof of concept for how future methods might control model behavior dynamically at runtime.