We propose a novel approach for training large language models (LLMs) to adhere to objectives defined within a latent embedding space.

Can neural networks systematically capture discrete, compositional task structure despite their continuous, distributed nature? The impressive capabilities of large-scale neural networks suggest that the answer to this question is yes. However, even for the most capable models, there are still frequent failure cases that raise doubts about their compositionality. Here, we seek to understand what it takes for a standard neural network to generalize over tasks that share compositional structure. We find that simply scaling data and model size leads to compositional generalization. We show that this holds across different task encodings as long as the training distribution sufficiently covers the task space. In line with this finding, we prove that standard multilayer perceptrons can approximate a general class of compositional task families to arbitrary precision using only a linear number of neurons with respect to the number of task modules. Finally, we uncover that if networks successfully compositionally generalize, the constituents of a task can be linearly decoded from their hidden activations. We show that this metric correlates with failures of text-to-image generation models to compose known concepts.

Speculative decoding is widely used in accelerating large language model (LLM) inference. In this work, we focus on the online draft model selection problem in speculative decoding. We design an algorithm that provably competes with the best draft model in hindsight for each query in terms of either the token acceptance probability or expected acceptance length. In particular, we show that we can accurately evaluate all draft models, instead of only the chosen model without incurring additional queries to the target model, which allows us to improve exponentially over the existing bandit-based approach as the number of draft models increases. Our approach is generically applicable with any speculative decoding methods (single draft, multi-drafts and draft-trees). Moreover, we design system-efficient versions of online learners and demonstrate that the overhead in computation and latency can be substantially reduced. We conduct extensive experiments on open-source LLMs and diverse datasets, demonstrating that our methods substantially outperform the state-of-the-art EAGLE3 and the BanditSpec baseline in a variety of domains where specialized domain-expert drafters are available, especially when long reasoning chains are required.

We train a linear attention transformer on millions of masked-block matrix completion tasks: each prompt is masked low-rank matrix whose missing block may be (i) a scalar prediction target or (ii) an unseen kernel slice of Nyström extrapolation. The model sees only input-output pairs and a mean-squared loss; it is given no normal equations, no handcrafted iterations, and no hint that the tasks are related. Surprisingly, after training, algebraic unrolling reveals the same parameter-free update rule across three distinct computational regimes (full visibility, rank-limited updates, and distributed computation). We prove that this rule achieves second-order convergence on full-batch problems, cuts distributed iteration complexity, and remains accurate with rank-limited attention. Thus, a transformer trained solely to patch missing blocks implicitly discovers a unified, resource-adaptive iterative solver spanning prediction, estimation, and Nyström extrapolation, highlighting a powerful capability of in-context learning.

Speculative decoding (SD) is a widely adopted approach for accelerating inference in large language models (LLMs), particularly when the draft and target models are well aligned. However, state-of-the-art SD methods typically rely on tightly coupled, self-attention-based Transformer decoders, often augmented with auxiliary pooling or fusion layers. This coupling makes them increasingly complex and harder to generalize across different models. We present Budget EAGLE (Beagle), the first, to our knowledge, cross-attention-based Transformer decoder SD model that achieves performance on par with leading self-attention SD models (EAGLE-v2) while eliminating the need for pooling or auxiliary components, simplifying the architecture, improving training efficiency, and maintaining stable memory usage during training-time simulation. To enable effective training of this novel architecture, we propose Two-Stage Block-Attention Training, a new method that achieves training stability and convergence efficiency in block-level attention scenarios. Extensive experiments across multiple LLMs and datasets show that Beagle achieves competitive inference speedups and higher training efficiency than EAGLE-v2, offering a strong alternative for architectures in speculative decoding.

Foundation models have recently emerged as powerful feature extractors in computational pathology, yet they typically omit mechanisms for leveraging the global spatial structure of tissues and the local contextual relationships among diagnostically relevan t regions -- key elements for understanding the tumor microenvironment. Multiple instance learning (MIL) remains an essential next step following foundation model, designing a framework to aggregate patch - level features into slide - level predictions. We presen t EAGLE - Net, a structure - preserving, attention - guided MIL architecture designed to augment prediction and interpretability. EAGLE - Net integrates multi - scale absolute spatial encoding to capture global tissue architecture, a top - K neighborhood - aware loss to focus attention on local microenvironments, and background suppression loss to minimize false positives. We benchmarked EAGLE - Net on large pan - cancer datasets, including three cancer types for classification (10,260 slides) and seven cancer types for surv ival prediction (4,172 slides), using three distinct histology foundation backbones (REMEDIES, Uni - V1, Uni2 - h). Across tasks, EAGLE - Net achieved up to 3% higher classification accuracy and the top concordance indices in 6 of 7 cancer types, producing smoot h, biologically coherent attention maps that aligned with expert annotations and highlighted invasive fronts, necrosis, and immune infiltration. These results position EAGLE - Net as a generalizable, interpretable framework that complements foundation models, enabling improved biomarker discovery, prognostic modeling, and clinical decision support.

Officials are to start using artificial intelligence to help estimate the age of asylum seekers who say they are children. Angela Eagle, the immigration minister, said on Tuesday the government would test technology that judges a person's age based on their facial features. It is the latest example of Labour ministers turning to AI to help solve problems with public services without spending significant amounts of money. The decision was announced on the same day that David Bolt, the chief inspector of borders and immigration, published a highly critical report into the haphazard way in which officials estimated the age of new arrivals. Eagle said in a written statement to parliament: "We have concluded that the most cost-effective option to pursue is likely to be facial age estimation, whereby AI technology – trained on millions of images where an individual's age is verifiable – is able to produce an age estimate with a known degree of accuracy for an individual whose age is unknown or disputed.

Temporal link prediction in dynamic graphs is a critical task with applications in diverse domains such as social networks, recommendation systems, and e-commerce platforms. While existing Temporal Graph Neural Networks (T-GNNs) have achieved notable success by leveraging complex architectures to model temporal and structural dependencies, they often suffer from scalability and efficiency challenges due to high computational overhead. In this paper, we propose EAGLE, a lightweight framework that integrates short-term temporal recency and long-term global structural patterns. EAGLE consists of a time-aware module that aggregates information from a node's most recent neighbors to reflect its immediate preferences, and a structure-aware module that leverages temporal personalized PageRank to capture the influence of globally important nodes. To balance these attributes, EAGLE employs an adaptive weighting mechanism to dynamically adjust their contributions based on data characteristics. Also, EAGLE eliminates the need for complex multi-hop message passing or memory-intensive mechanisms, enabling significant improvements in efficiency. Extensive experiments on seven real-world temporal graphs demonstrate that EAGLE consistently achieves superior performance against state-of-the-art T-GNNs in both effectiveness and efficiency, delivering more than a 50x speedup over effective transformer-based T-GNNs.

Accurate cancer survival prediction requires integration of diverse data modalities that reflect the complex interplay between imaging, clinical parameters, and textual reports. However, existing multimodal approaches suffer from simplistic fusion strategies, massive computational requirements, and lack of interpretability-critical barriers to clinical adoption. We present EAGLE (Efficient Alignment of Generalized Latent Embeddings), a novel deep learning framework that addresses these limitations through attention-based multimodal fusion with comprehensive attribution analysis. EAGLE introduces four key innovations: (1) dynamic cross-modal attention mechanisms that learn hierarchical relationships between modalities, (2) massive dimensionality reduction (99.96%) while maintaining predictive performance, (3) three complementary attribution methods providing patient-level interpretability, and (4) a unified pipeline enabling seamless adaptation across cancer types. We evaluated EAGLE on 911 patients across three distinct malignancies: glioblastoma (GBM, n=160), intraductal papillary mucinous neoplasms (IPMN, n=171), and non-small cell lung cancer (NSCLC, n=580). Patient-level analysis showed high-risk individuals relied more heavily on adverse imaging features, while low-risk patients demonstrated balanced modality contributions. Risk stratification identified clinically meaningful groups with 4-fold (GBM) to 5-fold (NSCLC) differences in median survival, directly informing treatment intensity decisions. By combining state-of-the-art performance with clinical interpretability, EAGLE bridges the gap between advanced AI capabilities and practical healthcare deployment, offering a scalable solution for multimodal survival prediction that enhances both prognostic accuracy and physician trust in automated predictions.

Unsupervised Domain Adaptation has been an efficient approach to transferring the semantic segmentation model across data distributions. Meanwhile, the recent Open-vocabulary Semantic Scene understanding based on large-scale vision language models is effective in open-set settings because it can learn diverse concepts and categories. However, these prior methods fail to generalize across different camera views due to the lack of cross-view geometric modeling. At present, there are limited studies analyzing cross-view learning. To address this problem, we introduce a novel Unsupervised Cross-view Adaptation Learning approach to modeling the geometric structural change across views in Semantic Scene Understanding.