Understanding the temporal dynamics of biological growth is critical across diverse fields such as microbiology, agriculture, and biodegradation research. Although vision-language models like Contrastive Language Image Pretraining (CLIP) have shown strong capabilities in joint visual-textual reasoning, their effectiveness in capturing temporal progression remains limited. To address this, we propose CLIPTime, a multimodal, multitask framework designed to predict both the developmental stage and the corresponding timestamp of fungal growth from image and text inputs. Built upon the CLIP architecture, our model learns joint visual-textual embeddings and enables time-aware inference without requiring explicit temporal input during testing. To facilitate training and evaluation, we introduce a synthetic fungal growth dataset annotated with aligned timestamps and categorical stage labels. CLIPTime jointly performs classification and regression, predicting discrete growth stages alongside continuous timestamps. We also propose custom evaluation metrics, including temporal accuracy and regression error, to assess the precision of time-aware predictions. Experimental results demonstrate that CLIPTime effectively models biological progression and produces interpretable, temporally grounded outputs, highlighting the potential of vision-language models in real-world biological monitoring applications.

Amodal completion, which is the process of inferring the full appearance of objects despite partial occlusions, is crucial for understanding complex human-object interactions (HOI) in computer vision and robotics. Existing methods, such as those that use pre-trained diffusion models, often struggle to generate plausible completions in dynamic scenarios because they have a limited understanding of HOI. To solve this problem, we've developed a new approach that uses physical prior knowledge along with a specialized multi-regional inpainting technique designed for HOI. By incorporating physical constraints from human topology and contact information, we define two distinct regions: the primary region, where occluded object parts are most likely to be, and the secondary region, where occlusions are less probable. Our multi-regional inpainting method uses customized denoising strategies across these regions within a diffusion model. This improves the accuracy and realism of the generated completions in both their shape and visual detail. Our experimental results show that our approach significantly outperforms existing methods in HOI scenarios, moving machine perception closer to a more human-like understanding of dynamic environments. We also show that our pipeline is robust even without ground-truth contact annotations, which broadens its applicability to tasks like 3D reconstruction and novel view/pose synthesis.

Loop invariants are essential for proving the correctness of programs with loops. Developing loop invariants is challenging, and fully automatic synthesis cannot be guaranteed for arbitrary programs. Some approaches have been proposed to synthesize loop invariants using symbolic techniques and more recently using neural approaches. These approaches are able to correctly synthesize loop invariants only for subsets of standard benchmarks. In this work, we investigate whether modern, reasoning-optimized large language models can do better. We integrate OpenAI's O1, O1-mini, and O3-mini into a tightly coupled generate-and-check pipeline with the Z3 SMT solver, using solver counterexamples to iteratively guide invariant refinement. We use Code2Inv benchmark, which provides C programs along with their formal preconditions and postconditions. On this benchmark of 133 tasks, our framework achieves 100% coverage (133 out of 133), outperforming the previous best of 107 out of 133, while requiring only 1-2 model proposals per instance and 14-55 seconds of wall-clock time. These results demonstrate that LLMs possess latent logical reasoning capabilities which can help automate loop invariant synthesis. While our experiments target C-specific programs, this approach should be generalizable to other imperative languages.

Recently, large language models (LLMs) have shown great promise in automating unit test generation, significantly reducing the manual effort required by developers. To effectively evaluate the capabilities of LLMs in this domain, it is crucial to have a well-designed benchmark that accurately reflects real-world scenarios and mitigates common pitfalls. Existing LLM test generation benchmarks are limited by two critical drawbacks: data contamination and structurally simple function code. As a result, we often cannot rely on the validity of scientific conclusions drawn from empirical studies using these limited benchmarks. The empirical evidence presented may be biased due to contamination and may fail to generalize beyond toy programs due to structural simplicity. To address these problems, we introduce ULT (UnLeakedTestbench), a new benchmark specifically designed for function-level unit test generation from real-world Python functions. ULT is constructed through a multi-stage curation process that ensures high cyclomatic complexity and mitigates test case contamination. With 3,909 carefully selected function-level tasks, ULT provides a more realistic and challenging evaluation of LLMs' test generation capabilities. We also provide PLT (PreLeakedTestbench), a pair benchmark of ULT with leaked tests designed to enable a controlled analysis of memorization versus reasoning in test generation. Our evaluation results demonstrate that ULT is significantly more challenging. For example, test cases generated by LLMs only achieve 41.32\%, 45.10\%, 30.22\%, and 40.21\% for accuracy, statement coverage, branch coverage, and mutation score on average for all LLMs, respectively. These results are substantially lower than the corresponding metrics on TestEval (91.79\%, 92.18\%, 82.04\%, and 49.69\%) and PLT (47.07\%, 55.13\%, 40.07\%, and 50.80\%).

Prompt tuning (PT), as an emerging resource-efficient fine-tuning paradigm, has showcased remarkable effectiveness in improving the task-specific transferability of vision-language models. This paper delves into a previously overlooked information asymmetry issue in PT, where the visual modality mostly conveys more context than the object-oriented textual modality. Correspondingly, coarsely aligning these two modalities could result in the biased attention, driving the model to merely focus on the context area. To address this, we propose DAPT, an effective PT framework based on an intuitive decouple-before-align concept. First, we propose to explicitly decouple the visual modality into the foreground and background representation via exploiting coarse-and-fine visual segmenting cues, and then both of these decoupled patterns are aligned with the original foreground texts and the hand-crafted background classes, thereby symmetrically strengthening the modal alignment. To further enhance the visual concentration, we propose a visual pull-push regularization tailored for the foreground-background patterns, directing the original visual representation towards unbiased attention on the region-of-interest object. We demonstrate the power of architecture-free DAPT through few-shot learning, base-to-novel generalization, and data-efficient learning, all of which yield superior performance across prevailing benchmarks. Our code will be released at https://github.com/Ferenas/DAPT.

Unmanned Aerial Vehicle (UAV) Vision-Language Navigation (VLN) aims to enable agents to accurately localize targets and plan flight paths in complex environments based on natural language instructions, with broad applications in intelligent inspection, disaster rescue, and urban monitoring. Recent progress in Vision-Language Models (VLMs) has provided strong semantic understanding for this task, while reinforcement learning (RL) has emerged as a promising post-training strategy to further improve generalization. However, existing RL methods often suffer from inefficient use of training data, slow convergence, and insufficient consideration of the difficulty variation among training samples, which limits further performance improvement. To address these challenges, we propose \textbf{Semantic-Aware Gaussian Curriculum Scheduling (SA-GCS)}, a novel training framework that systematically integrates Curriculum Learning (CL) into RL. SA-GCS employs a Semantic-Aware Difficulty Estimator (SA-DE) to quantify the complexity of training samples and a Gaussian Curriculum Scheduler (GCS) to dynamically adjust the sampling distribution, enabling a smooth progression from easy to challenging tasks. This design significantly improves training efficiency, accelerates convergence, and enhances overall model performance. Extensive experiments on the CityNav benchmark demonstrate that SA-GCS consistently outperforms strong baselines across all metrics, achieves faster and more stable convergence, and generalizes well across models of different scales, highlighting its robustness and scalability. The implementation of our approach is publicly available.

Small language models (SLMs) are inherently limited in knowledge-intensive tasks due to their constrained capacity. While test-time computation offers a path to enhanced performance, most approaches treat reasoning as a fixed or heuristic process. In this work, we propose a new paradigm: viewing the model's internal reasoning, delimited by and tags, as a dynamic task vector machine. Rather than treating the content inside these tags as a mere trace of thought, we interpret the generation process itself as a mechanism through which the model \textbf{constructs and refines its own task vectors} on the fly. We developed a method to optimize this dynamic task vector machine through RLVR and successfully trained an agentic web-search model. We present Lucy, a 1.7B-parameter SLM that leverages this dynamic reasoning mechanism with MCP integration to achieve 78.3% accuracy on the SimpleQA benchmark, performing on par with much larger models such as DeepSeek-V3. This demonstrates that small models can rival large ones when equipped with structured, self-constructed task reasoning.

Although large reasoning models (LRMs) have demonstrated impressive capabilities on complex tasks, recent studies reveal that these models frequently fulfill harmful user instructions, raising significant safety concerns. In this paper, we investigate the underlying cause of LRM safety risks and find that models already possess sufficient safety knowledge but fail to activate it during reasoning. Based on this insight, we propose R1-Act, a simple and efficient post-training method that explicitly triggers safety knowledge through a structured reasoning process. R1-Act achieves strong safety improvements while preserving reasoning performance, outperforming prior alignment methods. Notably, it requires only 1,000 training examples and 90 minutes of training on a single RTX A6000 GPU. Extensive experiments across multiple LRM backbones and sizes demonstrate the robustness, scalability, and practical efficiency of our approach.

Evaluating the performance of visual language models (VLMs) in graphic reasoning tasks has become an important research topic. However, VLMs still show obvious deficiencies in simulating human-level graphic reasoning capabilities, especially in complex graphic reasoning and abstract problem solving, which are less studied and existing studies only focus on simple graphics. To evaluate the performance of VLMs in complex graphic reasoning, we propose ReasonBench, the first evaluation benchmark focused on structured graphic reasoning tasks, which includes 1,613 questions from real-world intelligence tests. ReasonBench covers reasoning dimensions related to location, attribute, quantity, and multi-element tasks, providing a comprehensive evaluation of the performance of VLMs in spatial, relational, and abstract reasoning capabilities. We benchmark 11 mainstream VLMs (including closed-source and open-source models) and reveal significant limitations of current models. Based on these findings, we propose a dual optimization strategy: Diagrammatic Reasoning Chain (DiaCoT) enhances the interpretability of reasoning by decomposing layers, and ReasonTune enhances the task adaptability of model reasoning through training, all of which improves VLM performance by 33.5\%. All experimental data and code are in the repository: https://huggingface.co/datasets/cistine/ReasonBench.

Large language models (LLMs) excel across diverse natural language processing tasks but face resource demands and limited context windows. Although techniques like pruning, quantization, and token dropping can mitigate these issues, their efficacy in long-context scenarios and system evaluation remains underexplored. This paper systematically benchmarks these optimizations, characterizing memory usage, latency, and throughput, and studies how these methods impact the quality of text generation. We first analyze individual optimization methods for two LLM architectures supporting long context and then systematically evaluate combinations of these techniques to assess how this deeper analysis impacts performance metrics. We subsequently study the scalability of individual optimization methods on a larger variant with 70 billion-parameter model. Our novel insights reveal that naive combination inference optimization algorithms can adversely affect larger models due to compounded approximation errors, as compared to their smaller counterparts. Experiments show that relying solely on F1 obscures these effects by hiding precision-recall trade-offs in question answering tasks. By integrating system-level profiling with task-specific insights, this study helps LLM practitioners and researchers explore and balance efficiency, accuracy, and scalability across tasks and hardware configurations.