Developing models that can learn to reason is a notoriously challenging problem. We focus on reasoning in relational domains, where the use of Graph Neural Networks (GNNs) seems like a natural choice. However, previous work on reasoning with GNNs has shown that such models tend to fail when presented with test examples that require longer inference chains than those seen during training. This suggests that GNNs lack the ability to generalize from training examples in a systematic way, which would fundamentally limit their reasoning abilities. A common solution is to instead rely on neuro-symbolic methods, which are capable of reasoning in a systematic way by design. Unfortunately, the scalability of such methods is often limited and they tend to rely on overly strong assumptions, e.g.\ that queries can be answered by inspecting a single relational path. In this paper, we revisit the idea of reasoning with GNNs, showing that systematic generalization is possible as long as the right inductive bias is provided. In particular, we argue that node embeddings should be treated as epistemic states and that GNN should be parameterised accordingly. We propose a simple GNN architecture which is based on this view and show that it is capable of achieving state-of-the-art results. We furthermore introduce a benchmark which requires models to aggregate evidence from multiple relational paths. We show that existing neuro-symbolic approaches fail on this benchmark, whereas our considered GNN model learns to reason accurately.

The Maximal Update Parametrization ($\mu$P) aims to make the optimal hyperparameters (HPs) of a model independent of its size, allowing them to be swept using a cheap proxy model rather than the full-size target model. We present a new scheme, u-$\mu$P, which improves upon $\mu$P by combining it with Unit Scaling, a method for designing models that makes them easy to train in low-precision. The two techniques have a natural affinity: $\mu$P ensures that the scale of activations is independent of model size, and Unit Scaling ensures that activations, weights and gradients begin training with a scale of one. This synthesis opens the door to a simpler scheme, whose default values are near-optimal. This in turn facilitates a more efficient sweeping strategy, with u-$\mu$P models reaching a lower loss than comparable $\mu$P models and working out-of-the-box in FP8.

Predicting traffic accidents is the key to sustainable city management, which requires effective address of the dynamic and complex spatiotemporal characteristics of cities. Current data-driven models often struggle with data sparsity and typically overlook the integration of diverse urban data sources and the high-order dependencies within them. Additionally, they frequently rely on predefined topologies or weights, limiting their adaptability in spatiotemporal predictions. To address these issues, we introduce the Spatiotemporal Multiview Adaptive HyperGraph Learning (SMA-Hyper) model, a dynamic deep learning framework designed for traffic accident prediction. Building on previous research, this innovative model incorporates dual adaptive spatiotemporal graph learning mechanisms that enable high-order cross-regional learning through hypergraphs and dynamic adaptation to evolving urban data. It also utilises contrastive learning to enhance global and local data representations in sparse datasets and employs an advance attention mechanism to fuse multiple views of accident data and urban functional features, thereby enriching the contextual understanding of risk factors. Extensive testing on the London traffic accident dataset demonstrates that the SMA-Hyper model significantly outperforms baseline models across various temporal horizons and multistep outputs, affirming the effectiveness of its multiview fusion and adaptive learning strategies. The interpretability of the results further underscores its potential to improve urban traffic management and safety by leveraging complex spatiotemporal urban data, offering a scalable framework adaptable to diverse urban environments.

Large language models (LLMs) exhibit remarkable in-context learning (ICL) capabilities. However, the underlying working mechanism of ICL remains poorly understood. Recent research presents two conflicting views on ICL: One attributes it to LLMs' inherent ability of task recognition, deeming label correctness and shot numbers of demonstrations as not crucial; the other emphasizes the impact of similar examples in the demonstrations, stressing the need for label correctness and more shots. In this work, we provide a Two-Dimensional Coordinate System that unifies both views into a systematic framework. The framework explains the behavior of ICL through two orthogonal variables: whether LLMs can recognize the task and whether similar examples are presented in the demonstrations. We propose the peak inverse rank metric to detect the task recognition ability of LLMs and study LLMs' reactions to different definitions of similarity. Based on these, we conduct extensive experiments to elucidate how ICL functions across each quadrant on multiple representative classification tasks. Finally, we extend our analyses to generation tasks, showing that our coordinate system can also be used to interpret ICL for generation tasks effectively.

Israel strikes Yemen Houthis Dek: Israel launched its first ever strikes against Houthi rebels in Yemen just days after Jerusalem vowed revenge from a drone strike on Tel Aviv. A secret meeting between the U.S., Israel and the United Arab Emirates has been held to discuss a potential strategy on how the Gaza Strip will be governed once there is an end to the months-long war, Fox News confirmed Tuesday. The meeting, held in Abu Dhabi on Thursday, suggests that Israeli Prime Minister Benjamin Netanyahu may be looking to establish a plan for Gaza once the war is over, following repeated calls for a cease-fire. But details on the Thursday meeting – first reported by Axios – remain scarce, and it is unclear if options for ending the war were also discussed. Smoke and flames rise in the wake of an Israeli airstrike in Gaza on Nov. 2, 2023.

Meta AI launched in September 2023 using the Llama 2 learning language model. Nearly a year later, Meta has announced a new round of features for its AI assistant and a fresh LLM to support it: Llama 3.1. These updates include an expansion of who can access Meta AI. Thanks to the addition of Argentina, Chile, Colombia, Ecuador, Mexico, Peru and Cameroon, the assistant is now available in 22 countries. However, some of the new features are location or language-specific for the time being.

The Chinese Communist Party is a geopolitical cancer that will metastasize unless America can contain it with a once-in-a-generation investment in our national defense. Already, the CCP is actively colluding with Russia, prolonging Putin's war against Ukraine by blunting the impact of Western sanctions; it reaffirmed its support for Iran even after the deadly Oct. 7 attacks against Israel; and it has an explicit defense treaty with Kim Jung Un's North Korean dictatorship. To make matters even more dire, Chinese President Xi Jinping has instructed his People's Liberation Army to be ready to invade Taiwan by 2027. Chinese President Xi Jinping has instructed his People's Liberation Army to be ready to invade Taiwan by 2027. As George Washington counseled Congress in the nation's first ever inaugural address, "to be prepared for war is the most effectual means of preserving the peace."

We introduce Lifelong ICL, a problem setting that challenges long-context language models (LMs) to learn from a sequence of language tasks through in-context learning (ICL). We further introduce Task Haystack, an evaluation suite dedicated to assessing and diagnosing how long-context LMs utilizes contexts in Lifelong ICL. When given a task instruction and test inputs, long-context LMs are expected to leverage the relevant demonstrations in the Lifelong ICL prompt, avoid distraction and interference from other tasks, and achieve test accuracies that are not significantly worse than the Single-task ICL baseline. Task Haystack draws inspiration from the widely-adopted "needle-in-a-haystack" (NIAH) evaluation, but presents new and unique challenges. It demands that models (1) utilize the contexts with deeper understanding, rather than resorting to simple copying and pasting; (2) navigate through long streams of evolving topics and tasks, which closely approximates the complexities of real-world usage of long-context LMs. Additionally, Task Haystack inherits the controllability aspect of NIAH, providing model developers with tools and visualizations to identify model vulnerabilities effectively. We benchmark 12 long-context LMs using Task Haystack. We find that state-of-the-art closed models such as GPT-4o still struggle in this setting, failing 15% of the cases on average, while all open-weight models we evaluate further lack behind by a large margin, failing up to 61% of the cases. In our controlled analysis, we identify factors such as distraction and recency bias as contributors to these failure cases. Further, we observe declines in performance when task instructions are paraphrased at test time or when ICL demonstrations are repeated excessively, raising concerns about the robustness, instruction understanding, and true context utilization of current long-context LMs.

The full-dimensional (metric, Euclidean, least squares) multidimensional scaling stress loss function is combined with a quadratic external penalty function term. The trajectory of minimizers of stress for increasing values of the penalty parameter is then used to find (tentative) global minima for low-dimensional multidimensional scaling. This is illustrated with several one-dimensional and two-dimensional examples.

Abstract--We present a local-to-global and measure-theoretical approach to understanding datasets. The core idea is to form ulate a logifold structure and to interpret network models with restricted domains as local charts of datasets. In particul ar, this provides a mathematical foundation for ensemble machi ne learning. Our experiments demonstrate that logifolds can b e implemented to identify fuzzy domains and improve accuracy compared to taking average of model outputs. Additionally, we provide a theoretical example of a logifold, highlighting t he importance of restricting to domains of classifiers in an ens emble.