Understanding camera dynamics is a fundamental pillar of video spatial intelligence. However, existing multimodal models predominantly treat this task as a black-box classification, often confusing physically distinct motions by relying on superficial visual patterns rather than geometric cues. We present \textbf{CamReasoner}, a framework that reformulates camera movement understanding as a structured inference process to bridge the gap between perception and cinematic logic. Our approach centers on the Observation-Thinking-Answer (O-T-A) paradigm, which compels the model to articulate spatio-temporal observations and reason about motion patterns within an explicit reasoning block. To instill this capability, we construct a Large-scale Inference Trajectory Suite comprising 18k SFT reasoning chains and 38k RL feedback samples. To the best of our knowledge, \textbf{we are the first to employ RL for logical alignment in camera movement understanding}, ensuring motion inferences are grounded in structured visual reasoning rather than contextual guesswork. Built upon Qwen2.5-VL-7B, CamReasoner-7B improves binary classification accuracy from 73.8\% to 78.4\% and VQA accuracy from 60.9\% to 74.5\% over its backbone, consistently outperforming both proprietary and open-source baselines across multiple benchmarks.

We introduce Deep QP Safety Filter, a fully data-driven safety layer for black-box dynamical systems. Our method learns a Quadratic-Program (QP) safety filter without model knowledge by combining Hamilton-Jacobi (HJ) reachability with model-free learning. We construct contraction-based losses for both the safety value and its derivatives, and train two neural networks accordingly. In the exact setting, the learned critic converges to the viscosity solution (and its derivative), even for non-smooth values. Across diverse dynamical systems -- even including a hybrid system -- and multiple RL tasks, Deep QP Safety Filter substantially reduces pre-convergence failures while accelerating learning toward higher returns than strong baselines, offering a principled and practical route to safe, model-free control.

Large Language Models (LLMs) have demonstrated remarkable reasoning capabilities but often grapple with reliability challenges like hallucinations. While Knowledge Graphs (KGs) offer explicit grounding, existing paradigms of KG-augmented LLMs typically exhibit cognitive rigidity--applying homogeneous search strategies that render them vulnerable to instability under neighborhood noise and structural misalignment leading to reasoning stagnation. To address these challenges, we propose CoG, a training-free framework inspired by Dual-Process Theory that mimics the interplay between intuition and deliberation. First, functioning as the fast, intuitive process, the Relational Blueprint Guidance module leverages relational blueprints as interpretable soft structural constraints to rapidly stabilize the search direction against noise. Second, functioning as the prudent, analytical process, the Failure-Aware Refinement module intervenes upon encountering reasoning impasses. It triggers evidence-conditioned reflection and executes controlled backtracking to overcome reasoning stagnation. Experimental results on three benchmarks demonstrate that CoG significantly outperforms state-of-the-art approaches in both accuracy and efficiency.

In LLM-based text-to-SQL systems, unanswerable and underspecified user queries may generate not only incorrect text but also executable programs that yield misleading results or violate safety constraints, posing a major barrier to safe deployment. Existing refusal strategies for such queries either rely on output-level instruction following, which is brittle due to model hallucinations, or estimate output uncertainty, which adds complexity and overhead. To address this challenge, we formalize safe refusal in text-to-SQL systems as an answerability-gating problem and propose LatentRefusal, a latent-signal refusal mechanism that predicts query answerability from intermediate hidden activations of a large language model. We introduce the Tri-Residual Gated Encoder, a lightweight probing architecture, to suppress schema noise and amplify sparse, localized cues of question-schema mismatch that indicate unanswerability. Extensive empirical evaluations across diverse ambiguous and unanswerable settings, together with ablation studies and interpretability analyses, demonstrate the effectiveness of the proposed approach and show that LatentRefusal provides an attachable and efficient safety layer for text-to-SQL systems. Across four benchmarks, LatentRefusal improves average F1 to 88.5 percent on both backbones while adding approximately 2 milliseconds of probe overhead.

Language models perform well on grammatical agreement, but it is unclear whether this reflects rule-based generalization or memorization. We study this question for German definite singular articles, whose forms depend on gender and case. Using GRADIEND, a gradient-based interpretability method, we learn parameter update directions for gender-case specific article transitions. We find that updates learned for a specific gender-case article transition frequently affect unrelated gender-case settings, with substantial overlap among the most affected neurons across settings. These results argue against a strictly rule-based encoding of German definite articles, indicating that models at least partly rely on memorized associations rather than abstract grammatical rules.

Prior work on LLM-based privacy focuses on norm judgment over synthetic vignettes, rather than how people think about a specific data practice and formulate their opinions. We address this gap by designing PrivacyReasoner, an agent architecture grounded in three key ideas: (1) LLMs can detect subtle privacy cues in natural language and role-play human characteristics; (2) a user's ``privacy mind'' can be reconstructed from their real-world online comment history, distilling experiences, personality, and cultural orientations; and (3) a contextual filter can dynamically activate relevant privacy beliefs based on the contexts in a scenario. We evaluate PrivacyReasoner on real-world privacy discussions from Hacker News, using an LLM-as-a-Judge evaluator calibrated against an established privacy concern taxonomy to quantify reasoning faithfulness. PrivacyReasoner significantly outperforms baselines in predicting individual privacy concerns and generalizes across different domains, such as AI, e-commerce, and healthcare.

Critique-guided reinforcement learning (RL) has emerged as a powerful paradigm for training LLM agents by augmenting sparse outcome rewards with natural-language feedback. However, current methods often rely on static or offline critic models, which fail to adapt as the policy evolves. In on-policy RL, the agent's error patterns shift over time, causing stationary critics to become stale and providing feedback of diminishing utility. To address this, we introduce ECHO (Evolving Critic for Hindsight-Guided Optimization)}, a framework that jointly optimizes the policy and critic through a synchronized co-evolutionary loop. ECHO utilizes a cascaded rollout mechanism where the critic generates multiple diagnoses for an initial trajectory, followed by policy refinement to enable group-structured advantage estimation. We address the challenge of learning plateaus via a saturation-aware gain shaping objective, which rewards the critic for inducing incremental improvements in high-performing trajectories. By employing dual-track GRPO updates, ECHO ensures the critic's feedback stays synchronized with the evolving policy. Experimental results show that ECHO yields more stable training and higher long-horizon task success across open-world environments.

We introduce Olmo 3, a family of state-of-the-art, fully-open language models at the 7B and 32B parameter scales. Olmo 3 model construction targets long-context reasoning, function calling, coding, instruction following, general chat, and knowledge recall. This release includes the entire model flow, i.e., the full lifecycle of the family of models, including every stage, checkpoint, data point, and dependency used to build it. Our flagship model, Olmo 3 Think 32B, is the strongest fully-open thinking model released to-date.

Enhancing the temporal understanding of Multimodal Large Language Models (MLLMs) is essential for advancing long-form video analysis, enabling tasks such as temporal localization, action detection, and time-sensitive question answering. While reinforcement learning (RL) has recently been explored for improving temporal reasoning, existing approaches are often confined to limited task types and data, restricting their generalization across diverse temporal understanding scenarios. To address this challenge, we present TempR1, a temporal-aware multi-task reinforcement learning framework that systematically strengthens MLLMs' temporal comprehension. We curate a multi-task corpus that exposes the model to diverse temporal structures and semantics, and build upon the Group Relative Policy Optimization (GRPO) algorithm to achieve stable and effective cross-task optimization. Specifically, we categorize temporal tasks into three correspondence types between predicted intervals and ground-truth instances, and design tailored localization rewards for each, enabling TempR1 to capture fine-grained temporal dependencies and adapt to different temporal patterns. Extensive experiments demonstrate that TempR1 attains state-of-the-art performance across multiple benchmarks. Moreover, its joint optimization over complementary tasks yields a strong synergistic effect, enhancing both generalization and single-task performance, establishing a scalable and principled paradigm for temporal reasoning in MLLMs.

Open-vocabulary mobile manipulation (OVMM) requires robots to follow language instructions, navigate, and manipulate while updating their world representation under dynamic environmental changes. However, most prior approaches update their world representation only at discrete update points such as navigation targets, waypoints, or the end of an action step, leaving robots blind between updates and causing cascading failures: overlooked objects, late error detection, and delayed replanning. To address this limitation, we propose BINDER (Bridging INstant and DEliberative Reasoning), a dual process framework that decouples strategic planning from continuous environment monitoring. Specifically, BINDER integrates a Deliberative Response Module (DRM, a multimodal LLM for task planning) with an Instant Response Module (IRM, a VideoLLM for continuous monitoring). The two modules play complementary roles: the DRM performs strategic planning with structured 3D scene updates and guides what the IRM attends to, while the IRM analyzes video streams to update memory, correct ongoing actions, and trigger replanning when necessary. Through this bidirectional coordination, the modules address the trade off between maintaining awareness and avoiding costly updates, enabling robust adaptation under dynamic conditions. Evaluated in three real world environments with dynamic object placement, BINDER achieves substantially higher success and efficiency than SoTA baselines, demonstrating its effectiveness for real world deployment.

This paper introduces a novel architecture for trajectory-conditioned forecasting of future 3D scene occupancy. In contrast to methods that rely on variational autoencoders (VAEs) to generate discrete occupancy tokens, which inherently limit representational capacity, our approach predicts multi-frame future occupancy in an end-to-end manner directly from raw image features. Inspired by the success of attention-based transformer architectures in foundational vision and language models such as GPT and VGGT, we employ a sparse occupancy representation that bypasses the intermediate bird's eye view (BEV) projection and its explicit geometric priors. This design allows the transformer to capture spatiotemporal dependencies more effectively. By avoiding both the finite-capacity constraint of discrete tokenization and the structural limitations of BEV representations, our method achieves state-of-the-art performance on the nuScenes benchmark for 1-3 second occupancy forecasting, outperforming existing approaches by a significant margin. Furthermore, it demonstrates robust scene dynamics understanding, consistently delivering high accuracy under arbitrary future trajectory conditioning.

Vision-Language Navigation requires agents to act coherently over long horizons by understanding not only local visual context but also how far they have advanced within a multi-step instruction. However, recent Vision-Language-Action models focus on direct action prediction and earlier progress methods predict numeric achievements; both overlook the monotonic co-progression property of the observation and instruction sequences. Building on this insight, Progress-Think introduces semantic progress reasoning, predicting instruction-style progress from visual observations to enable more accurate navigation. To achieve this without expensive annotations, we propose a three-stage framework. In the initial stage, Self-Aligned Progress Pretraining bootstraps a reasoning module via a novel differentiable alignment between visual history and instruction prefixes. Then, Progress-Guided Policy Pretraining injects learned progress states into the navigation context, guiding the policy toward consistent actions. Finally, Progress-Policy Co-Finetuning jointly optimizes both modules with tailored progress-aware reinforcement objectives. Experiments on R2R-CE and RxR-CE show state-of-the-art success and efficiency, demonstrating that semantic progress yields a more consistent representation of navigation advancement.

Large language models often respond to ambiguous requests by implicitly committing to one interpretation, frustrating users and creating safety risks when that interpretation is wrong. We propose generating a single structured response that enumerates the different ways an ambiguous request can be interpreted, each coupled with a corresponding answer. Our models are trained with reinforcement learning using a dual reward objective: recall on ambiguous inputs to maximise coverage of valid interpretations, and precision on unambiguous ones to suppress spurious alternatives. Training requires only multiple valid answers per input as supervision, no clarification questions or explicit interpretations are needed. Experiments on conversational question answering and semantic parsing demonstrate that our method achieves higher coverage of valid answers than baseline approaches. Human evaluation confirms that predicted interpretations are meaningful and explain their corresponding answers. Our approach promotes transparency with explicit interpretations, achieves efficiency by requiring only one generation step, and supports downstream applications through its structured output format.

Retrieval-Augmented Generation (RAG) improves factuality but retrieving for every query often hurts quality while inflating tokens and latency. We propose Training-free Adaptive Retrieval Gating (TARG), a single-shot policy that decides when to retrieve using only a short, no-context draft from the base model. From the draft's prefix logits, TARG computes lightweight uncertainty scores-mean token entropy, a margin signal derived from the top-1/top-2 logit gap via a monotone link, or small-N variance across a handful of stochastic prefixes-and triggers retrieval only when the score exceeds a threshold. The gate is model-agnostic, adds only tens to hundreds of draft tokens, and requires no additional training or auxiliary heads. On five QA benchmarks spanning short-answer (NQ-Open, TriviaQA, PopQA), multi-hop (MuSiQue), and long-form (ASQA) tasks, TARG consistently pushes the accuracy-efficiency frontier: compared with Alway-RAG, TARG matches or improves EM/F1 while reducing retrieval by 70-90% and cutting end-to-end latency, and it remains close to Never-RAG in overhead. A central empirical finding is that under modern instruction-tuned LLMs the margin signal is a robust default (entropy compresses as backbones sharpen), with small-N variance offering a conservative, budget-first alternative. We provide ablations over gate type and prefix length and use a $Δ$-latency view to make budget trade-offs explicit.

Group Relative Policy Optimization (GRPO) has demonstrated wide adoption in the post-training of Large Language Models (LLMs). In GRPO, prompts are answered by the model and preferred behaviour is learnt via reinforcement learning. Owing to the small communication volume, GRPO is inherently suitable for decentralised training as the prompts can be concurrently answered by multiple nodes and these completions are exchanged in the form of strings. In this work, we explore the robustness of decentralised GRPO by presenting the first adversarial attacks and countermeasures. We present a diverse set of attacks where malicious nodes poison benign models by sharing their poisoned completions. We demonstrate these attacks on math and coding tasks and show that an adversary can achieve attack success rates of up to 100% in as few as 50 iterations. Moreover, to mitigate the attacks, we propose two defense mechanisms that check logit probabilities of completions or utilize an LLM judge to filter completions. The defenses prevent all but the DoS attack that causes unnecessarily lengthy but conceptually correct completions. The code of both attacks and defenses can be found at: https://github.com/gensyn-ai/HTTT.

Control barrier functions (CBFs) are a popular tool for safety certification of nonlinear dynamical control systems. Recently, CBFs represented as neural networks have shown great promise due to their expressiveness and applicability to a broad class of dynamics and safety constraints. However, verifying that a trained neural network is indeed a valid CBF is a computational bottleneck that limits the size of the networks that can be used. To overcome this limitation, we present a novel framework for verifying neural CBFs based on piecewise linear upper and lower bounds on the conditions required for a neural network to be a CBF. Our approach is rooted in linear bound propagation (LBP) for neural networks, which we extend to compute bounds on the gradients of the network. Combined with McCormick relaxation, we derive linear upper and lower bounds on the CBF conditions, thereby eliminating the need for computationally expensive verification procedures. Our approach applies to arbitrary control-affine systems and a broad range of nonlinear activation functions. To reduce conservatism, we develop a parallelizable refinement strategy that adaptively refines the regions over which these bounds are computed. Our approach scales to larger neural networks than state-of-the-art verification procedures for CBFs, as demonstrated by our numerical experiments.

In this paper we show that visual diffusion models can serve as effective geometric solvers: they can directly reason about geometric problems by working in pixel space. We first demonstrate this on the Inscribed Square Problem, a long-standing problem in geometry that asks whether every Jordan curve contains four points forming a square. We then extend the approach to two other well-known hard geometric problems: the Steiner Tree Problem and the Simple Polygon Problem. Our method treats each problem instance as an image and trains a standard visual diffusion model that transforms Gaussian noise into an image representing a valid approximate solution that closely matches the exact one. The model learns to transform noisy geometric structures into correct configurations, effectively recasting geometric reasoning as image generation. Unlike prior work that necessitates specialized architectures and domain-specific adaptations when applying diffusion to parametric geometric representations, we employ a standard visual diffusion model that operates on the visual representation of the problem. This simplicity highlights a surprising bridge between generative modeling and geometric problem solving. Beyond the specific problems studied here, our results point toward a broader paradigm: operating in image space provides a general and practical framework for approximating notoriously hard problems, and opens the door to tackling a far wider class of challenging geometric tasks.

Language models often struggle to follow multi-constraint instructions that are crucial for real-world applications. Existing reinforcement learning (RL) approaches suffer from dependency on external supervision and sparse reward signals from multi-constraint tasks. We propose a label-free self-supervised RL framework that eliminates dependency on external supervision by deriving reward signals directly from instructions and generating pseudo-labels for reward model training. Our approach introduces constraint decomposition strategies and efficient constraint-wise binary classification to address sparse reward challenges while maintaining computational efficiency. Experiments show that our approach generalizes well, achieving strong improvements across 3 in-domain and 5 out-of-domain datasets, including challenging agentic and multi-turn instruction following. The data and code are publicly available at https://github.com/Rainier-rq/verl-if

With the rapid adoption of diffusion models for visual content generation, proving authorship and protecting copyright have become critical. This challenge is particularly important when model owners keep their models private and may be unwilling or unable to handle authorship issues, making third-party verification essential. A natural solution is to embed watermarks for later verification. However, existing methods require access to model weights and rely on computationally heavy procedures, rendering them impractical and non-scalable. To address these challenges, we propose NoisePrints, a lightweight watermarking scheme that utilizes the random seed used to initialize the diffusion process as a proof of authorship without modifying the generation process. Our key observation is that the initial noise derived from a seed is highly correlated with the generated visual content. By incorporating a hash function into the noise sampling process, we further ensure that recovering a valid seed from the content is infeasible. We also show that sampling an alternative seed that passes verification is infeasible, and demonstrate the robustness of our method under various manipulations. Finally, we show how to use cryptographic zero-knowledge proofs to prove ownership without revealing the seed. By keeping the seed secret, we increase the difficulty of watermark removal. In our experiments, we validate NoisePrints on multiple state-of-the-art diffusion models for images and videos, demonstrating efficient verification using only the seed and output, without requiring access to model weights.

Trackers and video generators solve closely related problems: the former analyze motion, while the latter synthesize it. We show that this connection enables pretrained video diffusion models to perform zero-shot point tracking by simply prompting them to visually mark points as they move over time. We place a distinctively colored marker at the query point, then regenerate the rest of the video from an intermediate noise level. This propagates the marker across frames, tracing the point's trajectory. To ensure that the marker remains visible in this counterfactual generation, despite such markers being unlikely in natural videos, we use the unedited initial frame as a negative prompt. Through experiments with multiple image-conditioned video diffusion models, we find that these "emergent" tracks outperform those of prior zero-shot methods and persist through occlusions, often obtaining performance that is competitive with specialized self-supervised models.

While existing video editing methods excel with single subjects, they struggle in dense, multi-subject scenes, frequently suffering from attention dilution and mask boundary entanglement that cause attribute leakage and temporal instability. To address this, we propose ASTRA, a training-free framework for seamless, arbitrary-subject video editing. Without requiring model fine-tuning, ASTRA precisely manipulates multiple designated subjects while strictly preserving non-target regions. It achieves this via two core components: a prompt-guided multimodal alignment module that generates robust conditions to mitigate attention dilution, and a prior-based mask retargeting module that produces temporally coherent mask sequences to resolve boundary entanglement. Functioning as a versatile plug-and-play module, ASTRA seamlessly integrates with diverse mask-driven video generators. Extensive experiments on our newly constructed benchmark, MSVBench, demonstrate that ASTRA consistently outperforms state-of-the-art methods. Code, models, and data are available at https://github.com/XWH-A/ASTRA.

Federated inference, in the form of one-shot federated learning, edge ensembles, or federated ensembles, has emerged as an attractive solution to combine predictions from multiple models. This paradigm enables each model to remain local and proprietary while a central server queries them and aggregates predictions. Yet, the robustness of federated inference has been largely neglected, leaving them vulnerable to even simple attacks. To address this critical gap, we formalize the problem of robust federated inference and provide the first robustness analysis of this class of methods. Our analysis of averaging-based aggregators shows that the error of the aggregator is small either when the dissimilarity between honest responses is small or the margin between the two most probable classes is large. Moving beyond linear averaging, we show that problem of robust federated inference with non-linear aggregators can be cast as an adversarial machine learning problem. We then introduce an advanced technique using the DeepSet aggregation model, proposing a novel composition of adversarial training and test-time robust aggregation to robustify non-linear aggregators. Our composition yields significant improvements, surpassing existing robust aggregation methods by 4.7 - 22.2% in accuracy points across diverse benchmarks.

Large language models (LLMs), despite being safety-aligned, exhibit brittle refusal behaviors that can be circumvented by simple linguistic changes. As tense jailbreaking demonstrates that models refusing harmful requests often comply when rephrased in past tense, a critical generalization gap is revealed in current alignment methods whose underlying mechanisms are poorly understood. In this work, we introduce Activation-Scaling Guard (ASGuard), an insightful, mechanistically-informed framework that surgically mitigates this specific vulnerability. In the first step, we use circuit analysis to identify the specific attention heads causally linked to the targeted jailbreaking such as a tense-changing attack. Second, we train a precise, channel-wise scaling vector to recalibrate the activation of tense vulnerable heads. Lastly, we apply it into a "preventative fine-tuning", forcing the model to learn a more robust refusal mechanism. Across four LLMs, ASGuard effectively reduces the attack success rate of targeted jailbreaking while preserving general capabilities and minimizing over refusal, achieving a Pareto-optimal balance between safety and utility. Our findings underscore how adversarial suffixes suppress the propagation of the refusal-mediating direction, based on mechanistic analysis. Furthermore, our work showcases how a deep understanding of model internals can be leveraged to develop practical, efficient, and targeted methods for adjusting model behavior, charting a course for more reliable and interpretable AI safety.

The remarkable capabilities of modern large reasoning models are largely unlocked through post-training techniques such as supervised fine-tuning (SFT) and reinforcement learning (RL). However, the architectural mechanisms behind such improvements remain largely opaque. In this work, we use circuit analysis to demonstrate that post-training for complex reasoning sparks the emergence of novel, functionally specialized attention heads. These heads collectively support structured reasoning and computation. Our comparative analysis across various model families reveals that these emergent heads evolve differently under different training regimes. Distillation and SFT foster a cumulative addition of stable reasoning heads. In contrast, group relative policy optimization (GRPO) operates in a dynamic search mode: relatively few attention heads are iteratively activated, evaluated, and pruned, with their survival closely tracking fluctuations in the task reward signal. Furthermore, we find that controllable "think on/off" models do not possess dedicated "thinking" heads. Instead, turning off explicit reasoning triggers a broader-but less efficient-set of compensatory heads. Through ablation and qualitative analyses, we connect these circuit-level dynamics to a crucial performance trade-off: strengthened heads enable sophisticated problem-solving strategies for difficult problems but can also introduce "over-thinking" failure modes, such as calculation errors or logical loops on simpler tasks. These findings connect circuit-level dynamics to macro-level performance, identifying an inherent tension where complex reasoning comes at the cost of elementary computations. More broadly, our work points to future directions for training policy design, emphasizing the need to balance the development of effective reasoning strategies with the assurance of reliable, flawless execution.

Endoluminal endoscopic procedures are essential for diagnosing colorectal cancer and other severe conditions in the digestive tract, urogenital system, and airways. 3D reconstruction and novel-view synthesis from endoscopic images are promising tools for enhancing diagnosis. Moreover, integrating physiological deformations and interaction with the endoscope enables the development of simulation tools from real video data. However, constrained camera trajectories and view-dependent lighting create artifacts, leading to inaccurate or overfitted reconstructions. We present MedGS, a novel 3D reconstruction framework leveraging the unique property of endoscopic imaging, where a single light source is closely aligned with the camera. Our method separates light effects from tissue properties. MedGS enhances 3D Gaussian Splatting with a physically based relightable model. We boost the traditional light transport formulation with a specialized MLP capturing complex light-related effects while ensuring reduced artifacts and better generalization across novel views. MedGS achieves superior reconstruction quality compared to baseline methods on both public and in-house datasets. Unlike existing approaches, MedGS enables tissue modifications while preserving a physically accurate response to light, making it closer to real-world clinical use. Repository: https://github.com/gmum/MedGS

Replication of experimental results has been a challenge faced by many scientific disciplines, including the field of machine learning. Recent work on the theory of machine learning has formalized replicability as the demand that an algorithm produce identical outcomes when executed twice on different samples from the same distribution. Provably replicable algorithms are especially interesting for reinforcement learning (RL), where algorithms are known to be unstable in practice. While replicable algorithms exist for tabular RL settings, extending these guarantees to more practical function approximation settings has remained an open problem. In this work, we make progress by developing replicable methods for linear function approximation in RL. We first introduce two efficient algorithms for replicable random design regression and uncentered covariance estimation, each of independent interest. We then leverage these tools to provide the first provably efficient replicable RL algorithms for linear Markov decision processes in both the generative model and episodic settings. Finally, we evaluate our algorithms experimentally and show how they can inspire more consistent neural policies.

Large language models have demonstrated impressive capabilities across various domains. However, their general-purpose nature often limits their effectiveness in specialized fields such as energy, where deep technical expertise and precise domain knowledge are essential. In this paper, we introduce EnergyGPT, a domain-specialized language model tailored for the energy sector, developed by fine-tuning the LLaMA 3.1-8B model on a high-quality, curated corpus of energy-related texts. We consider two adaptation strategies: a full-parameter Supervised Fine-Tuning variant and a parameter-efficient LoRA-based variant that updates only a small fraction of the model parameters. We present a complete development pipeline, including data collection and curation, model fine-tuning, benchmark design and LLM-judge choice, evaluation, and deployment. Through this work, we demonstrate that our training strategy enables improvements in domain relevance and performance without the need for large-scale infrastructure. By evaluating the performance of both EnergyGPT variants using domain-specific question-answering benchmarks, our results show that the adapted models consistently outperform the base model in most energy-related language understanding and generation tasks, with the LoRA variant achieving competitive gains at significantly reduced training cost.

Accurate global crop type mapping supports agricultural monitoring and food security, yet remains limited by the scarcity of labeled data in many regions. A key challenge is enabling models trained in one geography to generalize reliably to others despite shifts in climate, phenology, and spectral characteristics. In this work, we show that geographic transfer in crop classification is primarily governed by the ability to learn invariant structure in multispectral time series. To systematically study this, we introduce CropGlobe, a globally distributed benchmark dataset of 300,000 samples spanning eight countries and five continents, and define progressively harder transfer settings from cross-country to cross-hemisphere. Across all settings, we find that simple spectral-temporal representations outperform both handcrafted features and modern geospatial foundation model embeddings. We propose CropNet, a lightweight convolutional architecture that jointly convolves across spectral and temporal dimensions to learn invariant crop signatures. Despite its simplicity, CropNet consistently outperforms larger transformer-based and foundation-model approaches under geographic domain shift. To further improve robustness to geographic variation, we introduce augmentations that simulate shifts in crop phenology and reflectance. Combined with CropNet, this yields substantial gains under large domain shifts. Our results demonstrate that inductive bias toward joint spectral-temporal structure is more critical for transfer than model scale or pretraining, pointing toward a scalable and data-efficient paradigm for worldwide agricultural mapping. Data and code are available at https://github.com/x-ytong/CropNet/.

Parameter-Efficient Fine-Tuning (PEFT) methods, such as Low-Rank Adaptation (LoRA), have significantly reduced the number of trainable parameters needed in fine-tuning large language models (LLMs). The developments of LoRA-style adapters have considered two main directions: (1) enhancing model expressivity with high-rank adapters, and (2) aiming for further parameter reduction, as exemplified by vector-based methods. However, these approaches come with a trade-off, as achieving the expressivity of high-rank weight updates typically comes at the cost of sacrificing the extreme parameter efficiency offered by vector-based techniques. To address this issue, we propose a vector-based random Tensor network for high-Rank Adaptation (TeRA), a novel PEFT method that achieves high-rank weight updates while retaining the parameter efficiency of vector-based PEFT adapters. This is achieved by parametrizing the tensorized weight update matrix as a Tucker-like tensor network (TN), whereby large randomly initialized factors are frozen and shared across layers, while only small layer-specific scaling vectors, corresponding to diagonal entries of factor matrices, are trained. Comprehensive experiments demonstrate that TeRA matches or even outperforms existing high-rank adapters, while requiring as few trainable parameters as vector-based methods. Theoretical analysis and ablation studies validate the effectiveness of the proposed TeRA method. The code is available at https://github.com/guyuxuan9/TeRA.

Achieving fully autonomous exploration and navigation remains a critical challenge in robotics, requiring integrated solutions for localisation, mapping, decision-making and motion planning. Existing approaches either rely on strict navigation rules lacking adaptability or on pre-training, which requires large datasets. These AI methods are often computationally intensive or based on static assumptions, limiting their adaptability in dynamic or unknown environments. This paper introduces a bio-inspired agent based on the Active Inference Framework (AIF), which unifies mapping, localisation, and adaptive decision-making for autonomous navigation, including exploration and goal-reaching. Our model creates and updates a topological map of the environment in real-time, planning goal-directed trajectories to explore or reach objectives without requiring pre-training. Key contributions include a probabilistic reasoning framework for interpretable navigation, robust adaptability to dynamic changes, and a modular ROS2 architecture compatible with existing navigation systems. Our method was tested in simulated and real-world environments. The agent successfully explores large-scale simulated environments and adapts to dynamic obstacles and drift, proving to be comparable to other exploration strategies such as Gbplanner, FAEL and Frontiers. This approach offers a scalable and transparent approach for navigating complex, unstructured environments.