Recent advances in large reasoning models (LRMs) have significantly enhanced long-chain reasoning capabilities over standard large language models (LLMs). However, LRMs often produce unnecessarily lengthy outputs even for simple queries, leading to inefficiencies or even accuracy degradation compared to LLMs. To address this, we propose CP-Router, a training-free, model-agnostic routing framework that dynamically selects between an LLM and an LRM, demonstrated with multiple-choice question answering (MCQA) prompts. The routing decision is guided by the prediction uncertainty estimates derived via Conformal Prediction (CP), which provides rigorous coverage guarantees. To improve uncertainty differentiation across inputs, we introduce Full and Binary Entropy (FBE), a novel entropy-based criterion that adaptively selects the appropriate CP threshold. Experiments across MCQA and QA benchmarks—including mathematics, logical reasoning, and Chinese chemistry—demonstrate that CP-Router efficiently reduces token usage while maintaining or even improving accuracy compared to using LRM alone. We further demonstrate the generality and robustness of CP-Router by extending it to diverse model pairings beyond the LLM–LRM setting.

Gas leaks pose significant risks to human health and the environment. Despite long-standing concerns, there are limited methods that can efficiently and accurately detect and segment leaks due to their concealed appearance and random shapes. In this paper, we propose a Fine-Grained Spatial-Temporal Perception (FGSTP) algorithm for gas leak segmentation. FGSTP captures critical motion clues across frames and integrates them with refined object features in an end-to-end network. Specifically, we first construct a correlation volume to capture motion information between consecutive frames. Then, the fine-grained perception progressively refines the object-level features using previous outputs. Finally, a decoder is employed to optimize boundary segmentation. Because there is no highly precise labeled dataset for gas leak segmentation, we manually label a gas leak video dataset, GasVid. Experimental results on GasVid demonstrate that our model excels in segmenting non-rigid objects such as gas leaks, generating the most accurate mask compared to other state-of-the-art (SOTA) models.

Predicting the performance of deep learning (DL) models, such as execution time and resource utilization, is crucial for Neural Architecture Search (NAS), DL cluster schedulers, and other technologies that advance deep learning. The representation of a model is the foundation for its performance prediction. However, existing methods cannot comprehensively represent diverse model configurations, resulting in unsatisfactory accuracy. To address this, we represent a model as a graph that includes the topology, along with node, edge, and global features, all of which are crucial for effectively capturing the performance of the model. Based on this representation, we propose PerfSeer, a novel predictor that uses a Graph Neural Network (GNN)-based performance prediction model, SeerNet. SeerNet fully leverages the topology and various features, while incorporating optimizations such as Synergistic Max-Mean aggregation (SynMM) and Global-Node Perspective Boost (GNPB) to more effectively capture the critical performance information, enabling it to predict the performance of models accurately. Furthermore, SeerNet can be extended to SeerNet-Multi by using Project Conflicting Gradients (PCGrad), enabling efficient simultaneous prediction of multiple performance metrics without significantly affecting accuracy. We constructed a dataset containing performance metrics for 53k+ model configurations, including execution time, memory usage, and Streaming Multiprocessor (SM) utilization during both training and inference. The evaluation results show that PerfSeer outperforms nn-Meter, Brp-NAS, and DIPPM.

Log-based anomaly detection is an important task in ensuring the stability and reliability of software systems. One of the key problems in this task is the lack of labeled logs. Existing works usually leverage large-scale labeled logs from mature systems to train an anomaly detection model of a target system based on the idea of transfer learning. However, these works still require a certain number of labeled logs from the target system. In this paper, we take a step forward and study a valuable yet underexplored setting: zero-label cross-system log-based anomaly detection, that is, no labeled logs are available in the target system. Specifically, we propose ZeroLog, a system-agnostic representation meta-learning method that enables cross-system log-based anomaly detection under zero-label conditions. To achieve this, we leverage unsupervised domain adaptation to perform adversarial training between the source and target domains, aiming to learn system-agnostic general feature representations. By employing meta-learning, the learned representations are further generalized to the target system without any target labels. Experimental results on three public log datasets from different systems show that ZeroLog reaches over $\mathbf{8 0 \%}$ F1-score without labels, comparable to state-of-the-art cross-system methods trained with labeled logs, and outperforms existing methods under zero-label conditions.

Abstract With the increment of the scale of data centers, device failures are inevitable, which requires high reliability of underlying network. Recursively defined structures can construct large-scale and high-performance data center networks (DCNs), where the investigation of reliability in this group of structures is insufficient as in the representative DPCell. In this paper, a comprehensive research on reliability of DPCell structure was carried out, including the diagnosability based on precise and pessimistic diagnostic strategies, $h$-extra connectivity, and $h$-extra conditional diagnosability. Let $r$ denote the number of layers in a DPCell structure. Let $k_{i}$ denote the degree used to connect to a vertex in the $i$th layer with $0 \leq i \leq r$. Let $h$ denote a scalar, where each connected component contains at least $h+1$ vertices in $G-F$. This research proves that (i) under the precise diagnosis strategy, ${DPCell}_{r}$ is $\sum _{i=0}^{r}k_{i}$-diagnosable for $r\geq 2$; (ii) under the pessimistic diagnosis strategy, ${DPCell}_{r}$ is $\left(\frac{k_{0}-1}{2}\right)/\left(\frac{k_{0}-1}{2}\right)$-diagnosable and $(\sum _{i=1}^{r}k_{i}-1)/(\sum _{i=1}^{r}k_{i}-1)$-diagnosable for $r\geq 0$; (iii) the $h$-extra connectivity $\kappa _{h}({DPCell}_{r})=(h+1)(\sum _{i=1}^{r}k_{i}-1)+k_{0}+1$ for $r\geq 1$ and $0\le h\le k_{0}$; and (iv) the $h$-extra conditional diagnosability $\widetilde{t_{h}^{P}}({DPCell}_{r})=(h+1)\sum _{i=1}^{r}k_{i}+k_{0}$ for $r\geq 2$ and $0\le h\le k_{0}$. Based on the research of connectivity and diagnosability, the reliability of DPCell structure can be evaluated quantitatively. This research results in a comprehensive understanding of the reliability of DPCell, which contributes to network performance improvement of recursively defined structures.

Abstract Remote monitoring systems has emerged to automate public services, where Internet of Things (IoT) devices are deployed to collect and relay streams of status updates to a remote control centre. The freshness of collected data is of critical importance to system performance, which can be quantified by the concept of age of information (AoI). To obtain timely perception of the monitored area, each sensor tends to sample status updates at the maximum frequency, which may lead to device drop‐outs due to limited battery capacity. Simultaneous wireless information and power transfer (SWIPT) is considered as a promising technology to prolong network lifetime. In this paper, a SWIPT‐enabled two‐hop relay network, where a relay node uses power splitting protocol to assist data transmissions from multiple source nodes to the destination is considered. The impact of power splitting coefficients, data arrival pattern and time‐varying channel conditions on the averaged age performance and expected lifetime of the system is theoretically analysed. Moreover, a lightweight relay strategy which selectively forward packets based on a reward function consisting of energy level at relay node and expected age reduction is developed. Simulation results show that the proposed strategy outperforms existing schemes in collecting status updates in a timely and energy‐efficient manner.