We use prompts to LLMs to act as language tools for two types of tasks in our work. The first being to798 read through and retrieve the relevant information from news articles to caption our image sequences,799 figures 6 and 7 The second being utilizing our captions to generate event specific question-answer800 pairs, figures 8 and 9.801 We conducted human validation on 144 events sampled across 15 disaster types to assess caption803 quality. Human evaluators were asked to classify each event as: (1) clear alignment between images,804 captions, and sources, (2) mismatch, or (3) inconclusive where imagery was insufficient to verify805 caption details. Overall results showed 65.3% clear alignment between images, captions, and sources,806 18.8% had mismatches, and 16.0% were inconclusive where imagery was insufficient to verify807 caption details. Excluding inconclusive cases, 77.7% of determinable events showed alignment,808 demonstrating reasonable caption quality for LLM-generated annotations.809

AF3 introduces: CMM (i) AF-Whisper, a unified audio encoder trainedPrevious SOTA (Closed Source) using a novel strategy for joint representation learning across all 3 modalities of speech, sound, and music; (ii) flexible, on-demand thinking, allowing the model to do chain-of-thought-type reasoning before answering; (iii) multi-turn, multiaudio chat; (iv) long audio understanding and reasoning (including speech) up MMSU to 10 minutes; and (v) voice-to-voice interaction. To enable these capabilities, (avg.)

Efficiently sampling from a complex probability distribution is a fundamental problem which has become increasingly pertinent in recent years with the rise of generative AI, as sophisticated sampling procedures from LLMs have been proposed to solve challenging reasoning problems. The efficacy of such sampling algorithms is limited, however, by the relationship between the LLM and the particular sampling task at hand, which has motivated the framework of test-time training (TTT). TTT works by updating a model's weights in response to partial generations and reward feedback received at inference time, thus adapting to the particular problem. In this work, we propose a formalization for TTT as the problem of producing a sample from a given probability measure $μ^\star$ belonging to a known class ${F}$ of distributions, given an oracle $\hat μ$ which yields approximate density estimates for $μ^\star$. This is closely related to the problem of reducing sampling to approximate counting studied in seminal works of Jerrum, Valiant & Vazirani (1986) and Jerrum & Sinclair (1989): namely, when ${F}$ is the class of all distributions, it coincides exactly with the aforementioned counting-to-sampling reduction. In this paper, we first show a quadratic lower bound on the query complexity of sampling from $μ^\star$ given query access to $\hat μ$ (for sufficiently large classes ${F}$), thus showing that the random walk approach proposed by Jerrum & Sinclair (1989) and refined by Hayes & Sinclair (2010), is optimal. This answers an open question posed by Hayes & Sinclair. We then show that this lower bound can be circumvented if the size of ${F}$ is bounded appropriately. As we discuss, this latter result can be viewed as an abstraction of TTT, and thus represents a starting point for the development of a principled theoretical framework for TTT.

AI generated predictions increasingly inform decision making in critical tasks, and therefore must be trustworthy. One widely used measure of trustworthiness is calibration, which requires that the predictions match the true frequencies and can be treated like real probabilities of a given outcome. However, defining calibration is subtle, and designing good measures of calibration error has been an active topic of recent research. The first goal is to find calibration measures that are actionable, meaning they can inform decision makers about their utility loss when predictions are treated as true probabilities, which is known as swap regret. The second goal is to find calibration measures that are testable, meaning that calibration error can be measured from a small sample of predictions and outcomes. Although these are very basic requirements, there is no existing calibration measure that fully satisfies both properties, and all existing measures relax actionability by bounding a weaker notion of swap regret, or relax testability by having suboptimal estimation error. We introduce a new calibration measure, Soft-Binned Calibration Decision Loss (SCDL), which we prove is fully actionable without weakening either requirement, and testable with nearly optimal error rate. In addition, SCDL satisfies other desired properties such as continuity and consistency. We also provide a set of experiments confirming that the theoretical advantages of SCDL compared to other measures lead to better performance in practice.

Despite extensive studies, the underlying reason as to why overparameterized neural networks can generalize remains elusive. Existing theory shows that common stochastic optimizers prefer flatter minimizers of the training loss, and thus a natural potential explanation is that flatness implies generalization. This work critically examines this explanation. Through theoretical and empirical investigation, we identify the following three scenarios for two-layer ReLU networks: (1) flatness provably implies generalization; (2) there exist non-generalizing flattest models and sharpness minimization algorithms fail to generalize poorly, and (3) perhaps most strikingly, there exist non-generalizing flattest models, but sharpness minimization algorithms still generalize. Our results suggest that the relationship between sharpness and generalization subtly depends on the data distributions and the model architectures and sharpness minimization algorithms do not only minimize sharpness to achieve better generalization. This calls for the search for other explanations for the generalization of over-parameterized neural networks.

However, foundational theoretical questions about this algorithm's privacy loss remain open--even in the seemingly simple setting of smooth convex losses over a bounded domain. Our main result resolves these questions: for a large range of parameters, we characterize the differential privacy up to a constant.

Chain-of-thought (CoT) rationales, which provide step-by-step reasoning to derive final answers, benefit LLMs in both inference and training. Incorporating rationales, either by generating them before answering during inference, or by placing them before or after the original answers during training - significantly improves model performance on mathematical, symbolic and commonsense reasoning tasks. However, most work focuses on the role of rationales in these reasoning tasks, overlooking their potential impact on other important tasks like natural language understanding (NLU) tasks. In this work, we raise the question: Can rationales similarly benefit NLU tasks? To conduct a systematic exploration, we construct NLURC, a comprehensive and high-quality NLU dataset collection with rationales, and develop various rationale-augmented methods. Through exploring the applicability of these methods on NLU tasks using the dataset, we uncover several potentially surprising findings: (1) CoT inference shifts from hindering NLU performance to surpassing direct label prediction as model size grows, indicating a positive correlation. (2) Most rationale-augmented training methods perform worse than label-only training, with one specially designed method consistently achieving improvements. (3) LLMs trained with rationales achieve significant performance gains on unseen NLU tasks, rivaling models ten times their size, while delivering interpretability on par with commercial LLMs.