Compositional Zero-Shot Learning (CZSL) aims to recognize unseen attribute-object compositions by leveraging knowledge of primitive concepts learned from seen compositions. Although recent works achieve impressive performance in CZSL by leveraging large vision-language models, they primarily rely on discriminative representations that may not explicitly preserve the structured relationships between primitive concepts and their compositions.
Natural hazard susceptibility mapping is often constrained by limited labeled data, reducing the generalizability of conventional machine learning and limiting the applicability of complex deep learning models. This study proposes SAGE (Spatially Augmented Graph Embeddings), a structurally informed feature-engineering framework that combines controlled noise-based data augmentation with neighborhood-based graph embeddings to improve prediction under data-scarce conditions.
The paper proposes a new dimension‑reduction strategy for transfer‑operator models of dynamical systems that focuses on scoring the σ‑algebra generated by coordinates rather than the operator’s spectral span. By using a χ²‑divergence criterion between embedded present and future states, the method guarantees that twice the intrinsic system dimension suffices to capture the full operator spectrum, even for systems with weakly interacting components that would otherwise require exponentially many modes. Experiments on benchmark systems show that this algebraic approach recovers masked components missed by rank‑based methods and enables accurate prediction of those components from few labels.
By Mark Kozdoba, Shie Mannor
BERTilda is an explainable framework for tracking topic lifecycles in longitudinal text streams. It discovers topics independently in each time window using an embedding‑based topic model, then links topics across adjacent windows via a temporal graph that uses both semantic similarity and a bidirectional coverage signal derived from tweet‑to‑topic attribution. The graph‑based rules identify continuations, splits, merges, disappearances, and unclear transitions, and the method achieves up to 87% agreement with human annotators on a gold‑standard subset.
By Cl\'audia Oliveira, \'Alvaro Figueira
The paper introduces a method for selecting a small, diverse subset from a large pool by addressing multiple, potentially conflicting notions of diversity. It formulates a fair multi‑view determinant selection problem that maximizes the weakest per‑view log determinant of a size‑k subset, smooths and relaxes the objective to the Stiefel manifold, and derives an adaptive self‑consistent‑field solver with damping and level shifting. The solver operates using only feature‑map products for each view and includes a rounding step via leverage‑score screening followed by fair local refinement.
By Richard Yi Da Xu
The paper introduces Debiased Inference with Multiple Imperfect Measurements (DMM), a framework that uses several error‑prone AI measurements to perform valid downstream statistical inference without requiring costly gold‑standard labels. By assuming conditional independence of the measurements given the true label and unit‑level features, DMM leverages CP decomposition and semiparametric theory to prove consistency and asymptotic normality of its estimator. Simulations demonstrate that DMM yields valid inference and can improve efficiency when additional imperfect measurements are available, and the authors provide diagnostics for the key independence assumption.
By Naoki Egami, Sooahn Shin
The paper introduces MIFR, a modality‑invariant and fair representation framework for skin disease classification that jointly processes clinical photographs and dermoscopic images using ViT‑based encoders. It employs a five‑component multi‑objective loss to balance classification accuracy, fairness across skin tones, class alignment, and modality invariance. Experiments on paired and external datasets demonstrate competitive predictive performance and fairness, with t‑SNE visualizations confirming alignment of embeddings from different modalities.
By Asonyu Senge Njih, Yvan Guifo Fodjo, Vianney Kengne Tchendji, Jerry Lacmou Zeutouo, Kerol Djoumessi
The paper presents a two‑stage multimodal framework for chest X‑ray interpretation that incorporates radiologist gaze data from the MIMIC‑Eye dataset. Stage 1 introduces a gaze‑token classifier that fuses image patches, bounding‑box masks, transcription embeddings, and fixation maps, and a curriculum‑scheduled loss that improves accuracy and spatial alignment, yielding a 4.4% AUC gain and 13.3% F1 improvement. Stage 2 translates classifier predictions into region‑specific diagnostic sentences using confidence‑weighted keywords, an expert dictionary, and a prompted large language model, boosting clinical‑term BERTScore and ROUGE over keyword baselines.
By Tanjim Islam Riju, Shuchismita Anwar, Saman Sarker Joy, Farig Sadeque, Swakkhar Shatabda
The paper introduces Temporally Decomposable Image Representations (TDIR), a learning algorithm that separates historical photographs into distinct date and content components using orthogonal subspaces. It establishes theoretical conditions for this decomposition, quantifies errors when conditions are partially met, and demonstrates that temporal and categorical subspaces naturally become orthogonal during joint optimization. TDIR supports transitive operations, allowing temporal information from one image to be transferred to another without label supervision, and is validated on composed image retrieval tasks that require both object content and time period specification.
By Adri\`a Molina Rodr\'iguez, Oriol Ramos Terrades, Josep Llad\'os Canet
MITRE‑SAGE is a multi‑agent retrieval‑augmented generation framework that combines semantic and structural cybersecurity knowledge to enhance large language model question‑answering. It decomposes tasks into query interpretation, evidence retrieval, and answer synthesis, supporting vulnerability assessment, threat profiling, and relationship extraction. The authors also introduce MITRE‑QA, a benchmark of 3,000 question‑answer pairs, and show that MITRE‑SAGE outperforms standalone LLMs and conventional RAG methods, with a lightweight configuration achieving top performance on most tasks.
By Ali Habibzadeh, Farid Feyzi, Reza Ebrahimi Atani
UMER is a Unified Multimodal Embedding and Ranking framework that combines contrastive embeddings with Pair‑Aware Discriminative Reasoning to improve universal multimodal retrieval. It replaces item‑wise reflection with pair‑wise comparison of query–candidate pairs, enabling explicit identification of matching and discrepancy evidence. A mutual distillation strategy transfers reliable pairwise preferences between the embedding and ranking components, and UMER achieves state‑of‑the‑art performance on the MMEB‑V2 benchmark while supporting budget‑adjustable inference.
By Libiao Chen, Xiyang Liu, Yanheng Wei, Tao Wang, Zhenyu Tang
The paper presents the CDSP (context-conditional deliberation signal pipeline), which transforms investment committee meeting transcripts into structured predictive features. CDSP segments transcripts into topical chunks, assigns asset‑class context labels via a large language model, maps financial keywords to a taxonomy, and adds sentiment polarity and mention frequency features. Using these engineered features on 48 monthly meetings, the best model—combining sentence embeddings with CDSP features—achieves 73% accuracy and a 0.73 F1 score, outperforming a simple stock‑choice baseline, though the improvement is not statistically significant.
By Vivek Batra, Kristin Chen, Sanjiv Das, Samuel Judge, Harshad Khadilkar, Sukrit Mittal, Amir Nasrollahzadeh, Daniel Ostrov, Jacob Sisk
The paper introduces MARCUS, a missing‑aware region representation model that treats missing data as a contextual urban signal rather than noise. MARCUS operates in three stages—Intra Learning, Inter Learning, and Fusion—to encode observed features, assess modality reliability, and generate final embeddings. Applied to rent prediction in Sydney and New York, MARCUS outperforms baselines, reducing MAE by 51.35% and 12.62% respectively, and additional experiments confirm its effectiveness.
By Chenya Huang, Bin Liang, Zhidong Li, Yuxi Lu, Kunqi Li, Justin Wang, Fang Chen
The paper introduces a computational framework that detects dynamic team‑process phases in collaborative virtual reality (VR) by analyzing timestamped dialogue. It uses late chunking, penalized Gaussian‑kernel change‑point detection, TF‑IDF, NMF, and a locally deployed large language model to identify semantic transitions and generate interpretable phase descriptions. The detected phases are aligned with interaction logs, demonstrating that transcript‑derived phases correspond to task‑action patterns and thus provide a transparent, transferable method for studying temporal changes in teamwork.
By Qing Huang, Jianing Zhang, Pooja Pol
The paper demonstrates that a single internal direction in modern language models—called the valence axis (V-axis)—captures how positive or negative a sentence feels. By using only nine emotion category names and 50 short narrative paragraphs per emotion, the authors identify this axis via principal component analysis of frozen encoder embeddings, achieving 93% of supervised performance on SST‑2 and strong correlations with human valence ratings across images, audio, and brain recordings. The method transfers across modalities without target‑modality labels, but works only for continuous attributes and is specific to certain model families.
By Yousef Radwan
StocksTalk is a voice‑enabled conversational agent that turns spoken financial screening requests into validated structured queries over real‑world market data. It integrates streaming speech recognition, retrieval‑augmented constraint extraction, schema‑grounded LLM‑based SQL generation, rule‑based validation, and human‑in‑the‑loop verification, exposing intermediate reasoning artifacts for user inspection. A benchmark of 150 spoken prompts shows that its retrieval grounding, constrained query generation, and interactive verification improve constraint extraction accuracy, SQL executability, logical consistency, and multi‑turn stability over baseline LLM approaches.
By Akshat Parmar, Vikranth Udandarao, Abhay Shakya, Tanmay Hire, Avinash Anand, Rajiv Ratn Shah, Daniel Wang Zhengkui
arXiv:2605. 05189v2 Announce Type: replace-cross Abstract: How many key-value associations can a $d\times d$ linear memory store?
By Nicholas Barnfield, Juno Kim, Eshaan Nichani, Jason D. Lee, Yue M. Lu
The paper introduces SIMPLE, a prior‑fitted multi‑view in‑context learner that learns a reusable, task‑conditioned inference procedure instead of a fixed fusion function. By generating synthetic task priors in embedding space, SIMPLE can handle diverse view configurations, class structures, and missingness patterns. Experiments on multi‑view and multi‑omics benchmarks show that a frozen SIMPLE model performs competitively, and lightweight adapter calibration further improves performance across most datasets.
By Jielong Lu, Zhihao Wu, Jiajun Yu, Zhaoliang Chen, Haishuai Wang
The paper introduces a federated active learning (FAL) approach that tackles data privacy and label scarcity by coordinating query selection across clients. In low-budget scenarios, it finds that homogeneous (IID) data actually requires stronger coordination to avoid redundant queries, while heterogeneous data naturally yields diversity—a reversal of the usual federated learning narrative. The authors propose a new framework that aligns client data in a shared embedding space via federated representation learning, enabling globally coordinated active selection while keeping annotations local, and demonstrate that this method outperforms existing FAL methods even with larger annotation budgets.
By Liam Mohr, Daphna Weinshall
The paper introduces TTSD‑FAR, a test‑time self‑distillation framework that adapts large video‑language models to missing‑modality scenarios in emotion recognition. A frozen teacher trained on complete modalities guides a low‑rank student, while Fisher‑Anchored Restoration monitors Fisher information to prevent drift and restore the student when distribution shifts occur. Experiments on MELD, DFEW, and BAH with up to 50% missing modalities show TTSD‑FAR consistently outperforms entropy‑based adaptation, retrieval‑augmented generation, and perplexity‑based generation, maintaining performance over long adaptation horizons.
By Muhammad Haseeb Aslam, Alessandro Koerich, Marco Pedersoli, Ali Etemad, Eric Granger