-- As space exploration advances, underground environments are becoming increasingly attractive due to their potential to provide shelter, easier access to resources, and enhanced scientific opportunities. Although such environments exist on Earth, they are often not easily accessible and do not accurately represent the diversity of underground environments found throughout the solar system. This paper presents PLUME, a procedural generation framework aimed at easily creating 3D underground environments. Its flexible structure allows for the continuous enhancement of various underground features, aligning with our expanding understanding of the solar system. The environments generated using PLUME can be used for AI training, evaluating robotics algorithms, 3D rendering, and facilitating rapid iteration on developed exploration algorithms. In this paper, it is demonstrated that PLUME has been used along with a robotic simulator . PLUME is open source and has been released on Github. To do planetary exploration, shelter will be an essential part to keep robots and humans protected from extreme temperatures, solar radiation, and micrometeorites. Existing subsurface structures such as caves or lava tubes are considered of high interest to create shelter.

Kleinberg and Mullainathan (2024) recently proposed an interesting model for language generation in the limit: Given a countable collection of languages, and an adversary enumerating the strings of some language $L$ from the collection, the objective is to generate new strings from the target language, such that all strings generated beyond some finite time are valid. Li, Raman and Tewari (2024) and Charikar and Pabbaraju (2024) showed strong non-uniform generation guarantees in this model, giving algorithms that generate new valid strings from $L$ after seeing a number of distinct input strings $t(L)$ that depends only on $L$ (and the collection), but not the enumeration order. However, for both these works, the language-wise generation times $t(L)$ of the algorithm can be strictly sub-optimal. In this work, we study Pareto-optimality of non-uniform language generation in the limit. We propose an algorithm, whose generation times $t^\star(L)$ are (almost) Pareto-optimal: any other algorithm whose generation time for some language $L$ is strictly smaller than $t^\star(L)$, must satisfy that its generation time for some other language $L'$ is strictly worse than $t^\star(L')$. Pareto-optimality is essentially the best that one can achieve for non-uniform generation. Our algorithmic framework conveniently adapts to further give Pareto-optimal non-uniform generation algorithms in the practically motivated settings of noisy as well as representative generation.

We introduce Prompt Curriculum Learning (PCL), a lightweight reinforcement learning (RL) algorithm that selects intermediate-difficulty prompts using a learned value model to post-train language models. Since post-training LLMs via RL remains sensitive to batching and prompt selection strategies, we first conduct a series of systematic experiments where we (1) determine the optimal training batch size that balances generation efficiency and gradient quality and (2) establish the importance of focusing on prompts of intermediate difficulty for the policy. We build upon these results to design PCL, which identifies prompts of intermediate difficulty for the current policy in an on-policy manner by using a value model that is concurrently updated based on the current policy. By focusing on informative prompts that yield high effective ratios, PCL achieves either the highest performance or requires significantly less time to reach comparable performance to its counterparts. Compared to rollout-based filtering methods, PCL avoids costly rollouts and achieves $12.1\times$ and $16.9\times$ faster speed on identifying intermediate-difficulty prompts when training on MATH and DeepScaleR, respectively. We further demonstrate that our value model accurately predicts prompt difficulty and allows PCL to focus on progressively more challenging prompts during RL. Our results present a new methodology that delivers improved tradeoff between upper-bound performance and efficiency for reasoning-focused RL.

This study examines the performance of today's open-source, locally hosted large-language models (LLMs) in handling complex competitive programming tasks with extended problem descriptions and contexts. Building on the original Framework for AI-driven Code Generation Evaluation (FACE), the authors retrofit the pipeline to work entirely offline through the Ollama runtime, collapsing FACE's sprawling per-problem directory tree into a handful of consolidated JSON files, and adding robust checkpointing so multi-day runs can resume after failures. The enhanced framework generates, submits, and records solutions for the full Kattis corpus of 3,589 problems across eight code-oriented models ranging from 6.7-9 billion parameters. The submission results show that the overall pass@1 accuracy is modest for the local models, with the best models performing at approximately half the acceptance rate of the proprietary models, Gemini 1.5 and ChatGPT-4. These findings expose a persistent gap between private, cost-controlled LLM deployments and state-of-the-art proprietary services, yet also highlight the rapid progress of open models and the practical benefits of an evaluation workflow that organizations can replicate on in-house hardware.

Small language models (SLMs) have the capability for text generation and may potentially be used to generate falsified texts online. This study evaluates 14 SLMs (1.7B-14B parameters) including LLaMA, Gemma, Phi, SmolLM, Mistral, and Granite families in generating perceived low and high quality fake news headlines when explicitly prompted, and whether they appear to be similar to real-world news headlines. Using controlled prompt engineering, 24,000 headlines were generated across low-quality and high-quality deceptive categories. Existing machine learning and deep learning-based news headline quality detectors were then applied against these SLM-generated fake news headlines. SLMs demonstrated high compliance rates with minimal ethical resistance, though there were some occasional exceptions. Headline quality detection using established DistilBERT and bagging classifier models showed that quality misclassification was common, with detection accuracies only ranging from 35.2% to 63.5%. These findings suggest the following: tested SLMs generally are compliant in generating falsified headlines, although there are slight variations in ethical restraints, and the generated headlines did not closely resemble existing primarily human-written content on the web, given the low quality classification accuracy.

--Generative AI is gaining increasing attention in software engineering, where testing remains an indispensable reliability mechanism. According to the widely adopted testing pyramid, unit tests constitute the majority of test cases and are often schematic, requiring minimal domain expertise. Automatically generating such tests under the supervision of software engineers can significantly enhance productivity during the development phase of the software lifecycle. This paper investigates the impact of code context and prompting strategies on the quality and adequacy of unit tests generated by various large language models (LLMs) across several families. The results show that including docstrings notably improves code adequacy, while further extending context to the full implementation yields definitely smaller gains. Notably, the chain-of-thought prompting strategy -- applied even to'reasoning' models -- achieves the best results, with up to 96.3% branch coverage, a 57% average mutation score, and near-perfect compilation success rate. Among the evaluated models, M5 (Gemini 2.5 Pro) demonstrated superior performance in both mutation score and branch coverage being still in top in terms of compilation success rate. ECENT years have brought significant advancements in artificial intelligence (AI), particularly in the areas of performance and productivity enhancement. However, AI -- and particularly large language models (LLMs) -- still suffer from several weaknesses. Among them, convincing but senseless content generation ('hallucination'), safety misalignment ('ethicality') [1], unfairness [2], and limited processing context are the most critical. In spite of these restrictions, and bearing in mind the limited and merely apparent creativity of LLMs [3], they have become versatile tools already widely used across a variety of domains (creative industries [4], entertainment, reporting, and software engineering [5] are just cases in point) for multiple tasks.

Reinforcement learning (RL) has become the core post-training technique for large language models (LLMs). RL for LLMs involves two stages: generation and training. The LLM first generates samples online, which are then used to derive rewards for training. The conventional view holds that the colocated architecture, where the two stages share resources via temporal multiplexing, outperforms the disaggregated architecture, in which dedicated resources are assigned to each stage. However, in real-world deployments, we observe that the colocated architecture suffers from resource coupling, where the two stages are constrained to use the same resources. This coupling compromises the scalability and cost-efficiency of colocated RL in large-scale training. In contrast, the disaggregated architecture allows for flexible resource allocation, supports heterogeneous training setups, and facilitates cross-datacenter deployment. StreamRL is designed with disaggregation from first principles and fully unlocks its potential by addressing two types of performance bottlenecks in existing disaggregated RL frameworks: pipeline bubbles, caused by stage dependencies, and skewness bubbles, resulting from long-tail output length distributions. To address pipeline bubbles, StreamRL breaks the traditional stage boundary in synchronous RL algorithms through stream generation and achieves full overlapping in asynchronous RL. To address skewness bubbles, StreamRL employs an output-length ranker model to identify long-tail samples and reduces generation time via skewness-aware dispatching and scheduling. Experiments show that StreamRL improves throughput by up to 2.66x compared to existing state-of-the-art systems, and improves cost-effectiveness by up to 1.33x in a heterogeneous, cross-datacenter setting.

The growing complexity of network traffic and demand for ultra-low latency communication require smarter packet traffic management. Existing Deep Learning-based queuing approaches struggle with dynamic network scenarios and demand high engineering effort. We propose AQM-LLM, distilling Large Language Models (LLMs) with few-shot learning, contextual understanding, and pattern recognition to improve Active Queue Management (AQM) [RFC 9330] with minimal manual effort. We consider a specific case where AQM is Low Latency, Low Loss, and Scalable Throughput (L4S) and our design of AQM-LLM builds on speculative decoding and reinforcement-based distilling of LLM by tackling congestion prevention in the L4S architecture using Explicit Congestion Notification (ECN) [RFC 9331] and periodic packet dropping. We develop a new open-source experimental platform by executing L4S-AQM on FreeBSD-14, providing interoperable modules to support LLM integration and facilitate IETF recognition through wider testing. Our extensive evaluations show L4S-LLM enhances queue management, prevents congestion, reduces latency, and boosts network performance, showcasing LLMs' adaptability and efficiency in uplifting AQM systems.

The evolutionary processes of complex systems contain critical information regarding their functional characteristics. The generation time of edges provides insights into the historical evolution of various networked complex systems, such as protein-protein interaction networks, ecosystems, and social networks. Recovering these evolutionary processes holds significant scientific value, including aiding in the interpretation of the evolution of protein-protein interaction networks. However, existing methods are capable of predicting the generation times of remaining edges given a partial temporal network but often perform poorly in cross-network prediction tasks. These methods frequently fail in edge generation time recovery tasks for static networks that lack timestamps. In this work, we adopt a comparative paradigm-based framework that fuses multiple networks for training, enabling cross-network learning of the relationship between network structure and edge generation times. Compared to separate training, this approach yields an average accuracy improvement of 16.98%. Furthermore, given the difficulty in collecting temporal networks, we propose a novel diffusion-model-based generation method to produce a large number of temporal networks. By combining real temporal networks with generated ones for training, we achieve an additional average accuracy improvement of 5.46% through joint training.