Spatial Context Improves the Integration of Text with Remote Sensing for Mapping Environmental Variables

Recent developments in natural language processing highlight text as an emerging data source for ecology. Textual resources carry unique information that can be used in complementarity with geospatial data sources, thus providing insights at the local scale into environmental conditions and properties hidden from more traditional data sources. Leveraging textual information in a spatial context presents several challenges. First, the contribution of textual data remains poorly defined in an ecological context, and it is unclear for which tasks it should be incorporated. Unlike ubiquitous satellite imagery or environmental covariates, the availability of textual data is sparse and irregular; its integration with geospatial data is not straightforward. In response to these challenges, this work proposes an attention-based approach that combines aerial imagery and geolocated text within a spatial neighbourhood, i.e. integrating contributions from several nearby observations. Our approach combines vision and text representations with a geolocation encoding, with an attention-based module that dynamically selects spatial neighbours that are useful for predictive tasks.The proposed approach is applied to the EcoWikiRS dataset, which combines high-resolution aerial imagery with sentences extracted from Wikipedia describing local environmental conditions across Switzerland. Our model is evaluated on the task of predicting 103 environmental variables from the SWECO25 data cube. Our approach consistently outperforms single-location or unimodal, i.e. image-only or text-only, baselines. When analysing variables by thematic groups, results show a significant improvement in performance for climatic, edaphic, population and land use/land cover variables, underscoring the benefit of including the spatial context when combining text and image data.

Building Envelope Inversion by Data-driven Interpretation of Ground Penetrating Radar

Ground-penetrating radar (GPR) combines depth resolution, non-destructive operation, and broad material sensitivity, yet it has seen limited use in diagnosing building envelopes. The compact geometry of wall assemblies, where reflections from closely spaced studs, sheathing, and cladding strongly overlap, has made systematic inversion difficult. Recent advances in data-driven interpretation provide an opportunity to revisit this challenge and assess whether machine learning can reliably extract structural information from such complex signals. Here, we develop a GPR-based inversion framework that decomposes wall diagnostics into classification tasks addressing vertical (stud presence) and lateral (wall-type) variations. Alongside model development, we implement multiple feature minimization strategies - including recursive elimination, agglomerative clustering, and L0-based sparsity - to promote fidelity and interpretability. Among these approaches, the L0-based sparse neural network (SparseNN) emerges as particularly effective: it exceeds Random Forest accuracy while relying on only a fraction of the input features, each linked to identifiable dielectric interfaces. SHAP analysis further confirms that the SparseNN learns reflection patterns consistent with physical layer boundaries. In summary, this framework establishes a foundation for physically interpretable and data-efficient inversion of wall assemblies using GPR radargrams. Although defect detection is not addressed here, the ability to reconstruct intact envelope structure and isolate features tied to key elements provides a necessary baseline for future inversion and anomaly-analysis tasks.

Radar-Based Identification of Individuals Using Heartbeat Features Extracted from Signal Amplitude and Phase

This study proposes a non-contact method for identifying individuals through the use of heartbeat features measured with millimeter-wave radar. Although complex-valued radar signal spectrograms are commonly used for this task, little attention has been paid to the choice of signal components, namely, whether to use amplitude, phase, or the complex signal itself. Although spectrograms can be constructed independently from amplitude or phase information, their respective contributions to identification accuracy remain unclear. To address this issue, we first evaluate identification performance using spectrograms derived separately from amplitude, phase, and complex signals. We then propose a feature fusion method that integrates these three representations to enhance identification accuracy. Experiments conducted with a 79-GHz radar system and involving six participants achieved an identification accuracy of 97.67%, demonstrating the effectiveness of the proposed component-wise analysis and integration approach.

EmbeddingRWKV: State-Centric Retrieval with Reusable States

Current Retrieval-Augmented Generation (RAG) systems typically employ a traditional two-stage pipeline: an embedding model for initial retrieval followed by a reranker for refinement. However, this paradigm suffers from significant inefficiency due to the lack of shared information between stages, leading to substantial redundant computation. To address this limitation, we propose \textbf{State-Centric Retrieval}, a unified retrieval paradigm that utilizes "states" as a bridge to connect embedding models and rerankers. First, we perform state representation learning by fine-tuning an RWKV-based LLM, transforming it into \textbf{EmbeddingRWKV}, a unified model that serves as both an embedding model and a state backbone for extracting compact, reusable states. Building upon these reusable states, we further design a state-based reranker to fully leverage precomputed information. During reranking, the model processes only query tokens, decoupling inference cost from document length and yielding a 5.4$\times$--44.8$\times$ speedup. Furthermore, we observe that retaining all intermediate layer states is unnecessary; with a uniform layer selection strategy, our model maintains 98.62\% of full-model performance using only 25\% of the layers. Extensive experiments demonstrate that State-Centric Retrieval achieves high-quality retrieval and reranking results while significantly enhancing overall system efficiency. Code is available at \href{https://github.com/howard-hou/EmbeddingRWKV}{our GitHub repository}.

Vector Quantized-Aided XL-MIMO CSI Feedback with Channel Adaptive Transmission

Efficient channel state information (CSI) feedback is critical for 6G extremely large-scale multiple-input multiple-output (XL-MIMO) systems to mitigate channel interference. However, the massive antenna scale imposes a severe burden on feedback overhead. Meanwhile, existing quantized feedback methods face dual challenges of limited quantization precision and insufficient channel robustness when compressing high-dimensional channel features into discrete symbols. To reduce these gaps, guided by the deep joint source-channel coding (DJSCC) framework, we propose a vector quantized (VQ)-aided scheme for CSI feedback in XL-MIMO systems considering the near-field effect, named VQ-DJSCC-F. Firstly, taking advantage of the sparsity of near-field channels in the polar-delay domain, we extract energy-concentrated features to reduce dimensionality. Then, we simultaneously design the Transformer and CNN (convolutional neural network) architectures as the backbones to hierarchically extract CSI features, followed by VQ modules projecting features into a discrete latent space. The entropy loss regularization in synergy with an exponential moving average (EMA) update strategy is introduced to maximize quantization precision. Furthermore, we develop an attention mechanism-driven channel adaptation module to mitigate the impact of wireless channel fading on the transmission of index sequences. Simulation results demonstrate that the proposed scheme achieves superior CSI reconstruction accuracy with lower feedback overheads under varying channel conditions.

Beyond Static Summarization: Proactive Memory Extraction for LLM Agents

Memory management is vital for LLM agents to handle long-term interaction and personalization. Most research focuses on how to organize and use memory summary, but often overlooks the initial memory extraction stage. In this paper, we argue that existing summary-based methods have two major limitations based on the recurrent processing theory. First, summarization is "ahead-of-time", acting as a blind "feed-forward" process that misses important details because it doesn't know future tasks. Second, extraction is usually "one-off", lacking a feedback loop to verify facts, which leads to the accumulation of information loss. To address these issues, we propose proactive memory extraction (namely ProMem). Unlike static summarization, ProMem treats extraction as an iterative cognitive process. We introduce a recurrent feedback loop where the agent uses self-questioning to actively probe the dialogue history. This mechanism allows the agent to recover missing information and correct errors. Our ProMem significantly improves the completeness of the extracted memory and QA accuracy. It also achieves a superior trade-off between extraction quality and token cost.

PII-VisBench: Evaluating Personally Identifiable Information Safety in Vision Language Models Along a Continuum of Visibility

Vision Language Models (VLMs) are increasingly integrated into privacy-critical domains, yet existing evaluations of personally identifiable information (PII) leakage largely treat privacy as a static extraction task and ignore how a subject's online presence--the volume of their data available online--influences privacy alignment. We introduce PII-VisBench, a novel benchmark containing 4000 unique probes designed to evaluate VLM safety through the continuum of online presence. The benchmark stratifies 200 subjects into four visibility categories: high, medium, low, and zero--based on the extent and nature of their information available online. We evaluate 18 open-source VLMs (0.3B-32B) based on two key metrics: percentage of PII probing queries refused (Refusal Rate) and the fraction of non-refusal responses flagged for containing PII (Conditional PII Disclosure Rate). Across models, we observe a consistent pattern: refusals increase and PII disclosures decrease (9.10% high to 5.34% low) as subject visibility drops. We identify that models are more likely to disclose PII for high-visibility subjects, alongside substantial model-family heterogeneity and PII-type disparities. Finally, paraphrasing and jailbreak-style prompts expose attack and model-dependent failures, motivating visibility-aware safety evaluation and training interventions.

QMAVIS: Long Video-Audio Understanding using Fusion of Large Multimodal Models

Large Multimodal Models (LMMs) for video-audio understanding have traditionally been evaluated only on shorter videos of a few minutes long. In this paper, we introduce QMAVIS (Q Team-Multimodal Audio Video Intelligent Sensemaking), a novel long video-audio understanding pipeline built through a late fusion of LMMs, Large Language Models, and speech recognition models. QMAVIS addresses the gap in long-form video analytics, particularly for longer videos of a few minutes to beyond an hour long, opening up new potential applications in sensemaking, video content analysis, embodied AI, etc. Quantitative experiments using QMAVIS demonstrated a 38.75% improvement over state-of-the-art video-audio LMMs like VideoLlaMA2 and InternVL2 on the VideoMME (with subtitles) dataset, which comprises long videos with audio information. Evaluations on other challenging video understanding datasets like PerceptionTest and EgoSchema saw up to 2% improvement, indicating competitive performance. Qualitative experiments also showed that QMAVIS is able to extract the nuances of different scenes in a long video audio content while understanding the overarching narrative. Ablation studies were also conducted to ascertain the impact of each component in the fusion pipeline.

Symbolic Regression for Shared Expressions: Introducing Partial Parameter Sharing

Symbolic Regression aims to find symbolic expressions that describe datasets. Due to better interpretability, it is a machine learning paradigm particularly powerful for scientific discovery. In recent years, several works have expanded the concept to allow the description of similar phenomena using a single expression with varying sets of parameters, thereby introducing categorical variables. Some previous works allow only "non-shared" (category-value-specific) parameters, and others also incorporate "shared" (category-value-agnostic) parameters. We expand upon those efforts by considering multiple categorical variables, and introducing intermediate levels of parameter sharing. With two categorical variables, an intermediate level of parameter sharing emerges, i.e., parameters which are shared across either category but change across the other. The new approach potentially decreases the number of parameters, while revealing additional information about the problem. Using a synthetic, fitting-only example, we test the limits of this setup in terms of data requirement reduction and transfer learning. As a real-world symbolic regression example, we demonstrate the benefits of the proposed approach on an astrophysics dataset used in a previous study, which considered only one categorical variable. We achieve a similar fit quality but require significantly fewer individual parameters, and extract additional information about the problem.

AutoTour: Automatic Photo Tour Guide with Smartphones and LLMs

We present AutoTour, a system that enhances user exploration by automatically generating fine-grained landmark annotations and descriptive narratives for photos captured by users. The key idea of AutoTour is to fuse visual features extracted from photos with nearby geospatial features queried from open matching databases. Unlike existing tour applications that rely on pre-defined content or proprietary datasets, AutoTour leverages open and extensible data sources to provide scalable and context-aware photo-based guidance. To achieve this, we design a training-free pipeline that first extracts and filters relevant geospatial features around the user's GPS location. It then detects major landmarks in user photos through VLM-based feature detection and projects them into the horizontal spatial plane. A geometric matching algorithm aligns photo features with corresponding geospatial entities based on their estimated distance and direction. The matched features are subsequently grounded and annotated directly on the original photo, accompanied by large language model-generated textual and audio descriptions to provide an informative, tour-like experience. We demonstrate that AutoTour can deliver rich, interpretable annotations for both iconic and lesser-known landmarks, enabling a new form of interactive, context-aware exploration that bridges visual perception and geospatial understanding.