Vision-Language Models (VLMs) like CLIP have advanced image representation through open-vocabulary semantic alignment. Yet, existing few-shot transfer learning methods largely overlook the intrinsic interdependencies between text and image embeddings, limiting their ability to fully transfer CLIP’s pretrained capabilities. To address this gap, we propose Hyperspherical Interpolation Variational Encoding (HIVE), a novel method for few-shot image classification. Our core idea is to shift away from directly training feature extraction capabilities for downstream tasks, and instead focus on exploring the semantic transformation relationships between upstream and downstream tasks. By modeling semantics from coarse to fine granularity, HIVE enables the transfer of original feature extraction and modality alignment capabilities to downstream tasks. Extensive experiments on eight established benchmarks, including CUB and EuroSAT, validate HIVE’s efficacy, achieving up to 46.2% and 80.0% improvements over the original CLIP in 1-shot and 16-shot classification tasks, respectively. Our work underscores the importance of preserving pretrained geometric constraints while exploiting semantic hierarchies for effective few-shot adaptation, providing a principled approach for vision-language model customization.

Large language models (LLMs) have made significant advancements in the field of natural language processing and are widely applied in tasks such as question answering, dialogue generation, and summarization. However, these models often face the "hallucination" problem, where they generate overly confident but inaccurate information when dealing with uncertain or unknown content, potentially leading to misleading or even harmful outcomes. To address this issue, this paper proposes a self-reflective rollback retrieval-augmented generation (SSR-RAG) method, which balances the accuracy and safety of generated content by integrating external knowledge retrieval and dynamic candidate set optimization. Specifically, we designed a chain retrieval strategy with a rollback mechanism that removes detrimental retrieval results from the candidate set if they negatively impact the final answer generation. Additionally, we introduced a query rewriting module to optimize user input queries and proposed a state evaluation function, PCF, which evaluates the quality of generated content by combining contextual perplexity and output confidence. Experimental results demonstrate that SSR-RAG achieves significant performance improvements on QA datasets such as MMLU, MMCU, and AGIEval, as well as on the safety validation dataset CValues, thereby validating its effectiveness and reliability.

Distributed Acoustic Sensing (DAS) can capture spatio-temporal vibration images of vehicles on expressways, which can be utilized for traffic monitoring. Compared to ubiquitously deployed cameras, DAS traffic monitoring offers advantages such as full coverage, resistance to environmental interference, low computational requirements, and cost-effectiveness. However, real-world complexities result in challenges for DAS traffic images, including low signal-to-noise ratio, signal missing, and uneven intensity. As DAS traffic applications are still in their early stages, effective solutions to these challenges are yet to be developed. This paper proposes a new deep learning method named DAS High Speed Traffic Trajectory (DAS-HTT) network, which contributes threefold: (i) Multi-Scale Context Extraction Module (MSCE) effectively enlarges the receptive field to capture long-range contextual information comprehensively; (ii) Stripe Convolution Decoder (SCD) acquires remote information along four directions, preventing irrelevant region interference in feature learning; (iii) Hierarchically Hough Transform Fusion Decoder (HHTFD) introduces the structural information of trajectory linearity, reducing the reliance on label data while enhancing trajectory continuity. We conducted experiments on an operating expressway, demonstrating that DAS-HTT outperforms existing methods across seven metrics, providing trajectories that are more consistent with ground truthes.

This paper develops an event-triggered sliding mode control (ETSMC) strategy for partially unknown disturbed systems via adaptive dynamic programming, where the system matrix is considered to be unknown to the designer. Both the sliding function and ETSMC scheme are constructed without using system matrix. An input-based event-triggered mechanism is introduced between the plant and the sliding mode controller to reduce the communication frequency. Compared with existing results on ETSMC, the proposed event-triggered mechanism can guarantee the reachability to the ideal sliding surface$s(t)=0$and thus the external disturbances can be eliminated completely. An online policy iteration algorithm is formulated to implement the partial-model-free ETSMC strategy. It is proven that in all policy iteration steps, the reachability of the prescribed sliding surface and the optimal control performance of the sliding mode dynamics are ensured simultaneously by the proposed online updated ETSMC scheme as well as the Zeno phenomenon of the proposed event generator can be excluded. Finally, the effectiveness and the applicability of the proposed ETSMC scheme are illustrated by a numerical example and a real experiment on the permanent magnet synchronous motor speed regulation system.Note to Practitioners—ETSMC is an effective robust control strategy for the practical networked control systems that can compensate the matched disturbances in the plant as well as reduce the information transmission frequency between the plant and controller. However, the design of the existing ETSMC strategies depends on the completely known system dynamics, which is difficult or expensive to be obtained in many engineering applications. Meanwhile, the ideal sliding motion cannot be attained by the existing ETSMC approaches so that the disturbance rejection performance is degraded unsatisfactorily. To address these concerns, this paper develops a novel partial-model-free ETSMC strategy based on adaptive dynamic programming for disturbed systems to achieve the optimal control performance without using the system matrix. The reachability of the ideal sliding surface and the exclusion of the Zeno behavior as well as the convergence of the online policy iteration algorithm are analyzed theoretically. The engineering applicability of the novel ETSMC scheme is verified in the speed regulation problem of the permanent magnet synchronous motor.

Data-driven seismic inversion has achieved certain advancements. However, these methods often require a large number of expensive well logs, limiting their application only to mature or synthetic data. This article presents ContrasInver, a method that achieves seismic inversion using as few as two or three well logs, significantly reducing the current requirements. In ContrasInver, two key innovations are proposed to address the challenges of applying semi-supervised learning to regression tasks with ultra-sparse labels: 1) the region-growing training (RGT) strategy leverages the inherent continuity of seismic data, effectively propagating accuracy from closer to more distant regions based on the proximity of well logs. To realize this concept, a multidimensional sample generation (MSG) method is also proposed that produces a large number of diverse samples from a single well, while establishing lateral continuity within the seismic data; 2) the impedance vectorization projection (IVP) vectorizes impedance values and performs semi-supervised learning in a compressed space. The Jacobian matrix derived from this space can filter out some outlier components in pseudo-label vectors, thereby solving the value confusion issue in semi-supervised regression learning. In the experiments, ContrasInver achieved state-of-the-art performance on the synthetic SEAM I data. In the field data with two or three well logs, only the methods based on the components proposed in this article were able to achieve reasonable results. It is the first data-driven approach yielding reliable results on the Netherlands F3 and Delft, using only three and two well logs, respectively.

Cross-domain lithology identification (CDLI) is a common case in lithology identification, which aims to train a machine learning model using the logging data of an interpreted well to predict the lithology of another uninterpreted well. Compared with the general lithology identification problem, the CDLI problem is more challenging for two reasons: the data distribution shift between the wells, and the expensive label acquisition on the uninterpreted well. To tackle these issues, we propose a novel framework that embeds active learning (AL) and domain adaptation into lithology identification. The proposed framework is composed of two components: an AL algorithm that selects the most uncertain and diverse target samples to query their real labels, and a source reweighting method that leverages the target labels to reduce data distribution discrepancy. Experimental results on two real-world data sets demonstrate that the proposed method can more effectively suppress the performance degradation caused by the data distribution shift than the baselines, with fewer target label queries.

The recent years have witnessed a great success of artificial intelligence applications in geological prospecting, so that the traditional manual work, which is time-consuming and labor-intensive, could be accomplished automatically or at least in a human–machine cooperation way. This letter presents a first attempt in proposing an automatic way to predict the cross-well sandstone that plays a crucial role in formation characterization and reservoir exploration. Such a two-stage framework is composed of i) a convolution neural network (CNN)-based coarse prediction module and ii) a geological experience-based error correction (EC) module. Experiments demonstrate that our proposed module can achieve comparable accuracy as experts.

Modern data centers host diverse applications, which generate a mix of short flows with stringent latency requirement and long flows requiring large sustained throughput. To solve the problem of resource competition between the mixed flows, we propose an adaptive traffic isolation scheme APS. Based on the packet spraying scheme in the multipath transmission, APS dynamically separates long flows from short ones on different paths to provide the low latency for the short flows. Meanwhile, to resolve the out-of-order problem, APS limits the long flows to a few paths with Equal Cost Multi Path (ECMP). Experimental results of NS2 simulation and testbed implementation show that, APS reduces the average completion time for short flows by up to 60 percent and increases the throughputs for long flows by about 1.68x over the state-of-the-art multipath transmission schemes.

Lithology identification plays an essential role in formation characterization and reservoir exploration. As an emerging technology, intelligent logging lithology identification has received great attention recently, which aims to infer the lithology type through the well-logging curves using machine-learning methods. However, the model trained on the interpreted logging data is not effective in predicting new exploration well due to the data distribution discrepancy. In this article, we aim to train a lithology identification model for the target well using a large amount of source-labeled logging data and a small amount of target-labeled data. The challenges of this task lie in three aspects: 1) the distribution misalignment; 2) the data divergence; and 3) the cost limitation. To solve these challenges, we propose a novel active adaptation for logging lithology identification (AALLI) framework that combines active learning (AL) and domain adaptation (DA). The contributions of this article are three-fold: 1) the domain-discrepancy problem in intelligent logging lithology identification is first investigated in this article, and a novel framework that incorporates AL and DA into lithology identification is proposed to handle the problem; 2) we design a discrepancy-based AL and pseudolabeling (PL) module and an instance importance weighting module to query the most uncertain target information and retain the most confident source information, which solves the challenges of cost limitation and distribution misalignment; and 3) we develop a reliability detecting module to improve the reliability of target pseudolabels, which, together with the discrepancy-based AL and PL module, solves the challenge of data divergence. Extensive experiments on three real-world well-logging datasets demonstrate the effectiveness of the proposed method compared to the baselines.

Subsurface lithofacies classification refers to the way of establishing a classifier on the interpreted well logging data to predict the lithofacies types corresponding to the uninterpreted ones. Such a task has become a research focus by applying machine learning technologies under the assumption of independent and identical distribution. However, due to the differences in, such as sedimentary environments, reservoir heterogeneity, and logging equipments, the same lithofacies might exhibit different logging characteristics between two different wells or even two different strata in one well. Therefore, motivated by the data drift issue and inspired by the domain adaptation methods, we propose the robust unilateral alignment (RUA) for lithofacies classification. The characteristics of the proposed RUA are as follows: 1) the projected maximum mean discrepancy (PMMD) is designed to reduce the marginal and conditional distribution discrepancy; 2) the random data mapping and target domain information preserving constraint is adopted to embody the data transformation model; and 3) the weighting mechanism and risk-aware constraint are introduced to solve the class imbalance and iterative risk problems. The experiments conducted on the data sets from Jiyang Depression, Bohai Bay Basin, verify the superior performances in accuracy and stability over the existing work.

Identifying borehole lithofacies through geop- hysical loggings is a fundamental task in petroleum exploration industry. Recent interdisciplinary studies have demonstrated the feasibility of applying machine learning to lithofacies identification. Most of these studies establish a mapping from the logging values at one depth point to the lithofacies type. However, due to the intrinsic properties of geophysical loggings, the logging shape should be taken into consideration, apart from the absolute values. In this article, we present the attempt to predict the lithofacies by feeding logging segments, and for the first time model the logging lithofacies identification problem as 1-D semantic segmentation. Such a logging segmentation task is challenging due to two reasons, strong spatial heterogeneity of lithofacies subsurface distribution and the explicit physical significance of geophysical loggings. To solve these challenges, we propose a novel geophysical logging segmentation network entitled SegLog. Specifically, we develop a global statistics pooling subnetwork and a statistics fusion subnetwork to generate statistical embeddings of geophysical loggings. Based on these statistical embeddings, we design a pixel-enhanced convolutional subnetwork to learn the microdetailed features, indicated by pixel-level logging values. These features are fused with the macrosemantic features extracted by a backbone U-Net to constitute the representations that can simultaneously describe the logging spatial correlation and pixel specificity. Experimental results on two logging datasets from the Jiyang Depression verify the effectiveness of our modeling strategy and its state-of-the-art performance on the lithofacies identification problem.

In this letter, considering the lack of core and drilling cuttings, an interpretable semisupervised classification method (ISSCM) under multiple smoothness assumptions is proposed and applied to lithology identification. The contribution is threefold. First, the novel semisupervised learning algorithm is developed based on the decision tree, the interpretability of which is highly beneficial to solve risk-aware problems. Second, both smoothness in the feature space and depth is utilized to generate pseudo-labels for the unlabelled data by using label propagation. Third, an algorithm to approximate the optimal affinity matrix is added to avoid degradation rendered by inappropriate manual settings under multiple smoothness assumptions. All these contributions could yield a classification model that is interpretable, accurate, and insusceptible to imprecise empirical settings. In the experiment, the proposed method is applied to lithology identification and verified by real-world data.

Modern datacenter topologies typically are multi-rooted trees consisting of multiple paths between any given pair of hosts. Recent load balancing designs focus on making full use of available parallel paths to provide high bisection bandwidth. However, they are agnostic to the mixed traffic generated by diverse applications in data centers and respectively use the same granularity in rerouting flows regardless of the flow type. Therefore, the short flows suffer the long-tailed queueing delay and reordering problems, while the throughputs of long flows are also degraded dramatically due to low link utilization and packet reordering under the non-adaptive granularity. To solve these problems, we design a traffic-aware load balancing (TLB) scheme to adopt different rerouting granularities for two kinds of flows. Specifically, TLB adaptively adjusts the switching granularity of long flows according to the load strength of short ones. Under the heavy load of short flows, the long flows use large switching granularity to help short ones obtain more opportunities in choosing short queues to complete quickly. When the load strength of short flows is low, the long flows switch paths more flexibly with small switching granularity to achieve high throughput. TLB is deployed at the switch, without any modifications on the end-hosts. The experimental results of NS2 simulations and Mininet implementation show that TLB significantly reduces the average flow completion time (AFCT) of short flows by ~15%-40% over the state-of-the-art load balancing schemes and achieves the high throughput for long flows.

Modern data-center applications generate a diverse mix of short and long flows with different performance requirements and weaknesses. The short flows are typically delay-sensitive but to suffer the head-of-line blocking and out-of-order problems. Recent solutions prioritize the short flows to meet their latency requirements, while damaging the throughput-sensitive long flows. To solve these problems, we design a Coding-based Adaptive Packet Spraying (CAPS) that effectively mitigates the negative impact of short and long flows on each other. To exploit the availability of multiple paths and avoid the head-of-line blocking, CAPS spreads the packets of short flows to all paths, while the long flows are limited to a few paths with Equal Cost Multi Path (ECMP). Meanwhile, to resolve the out-of-order problem with low overhead, CAPS encodes the short flows using forward error correction (FEC) technology and adjusts the coding redundancy according to the blocking probability. Moreover, since the coding efficiency decreases when the coding unit is too small or large, we demonstrate how to obtain the optimal size of coding unit. The coding layer is deployed between the TCP and IP layers, without any modifications on the existing TCP/IP protocols. The test results of NS2 simulation and small-scale testbed experiments show that CAPS significantly reduces the average flow completion time of short flows by ~30%-70% over the state-of-the-art multipath transmission schemes and achieves the high throughput for long flows with negligible traffic overhead.

This paper presents a novel indoor localization method for skid-steering mobile robot by fusing the readings from encoder, gyroscope, and magnetometer which can be read as an enhanced dead-reckoning localization method. Compared with the traditional dead-reckoning localization method implemented by encoder only, the accuracy and reliability can be improved significantly in spite of the price of slightly higher cost in digital devices. The proposed strategy consists mainly of an orientation algorithm and a localization algorithm. First, realizing that gyroscope is barely affected by magnetic field and magnetometer-based orientation has no cumulative error, a novel orientation algorithm, based on the self-tuning Kalman filter coupled with a gross error recognizer, is developed. This orientation algorithm can be applied to determine the robot heading angle in the situation with abundant ferromagnetic materials. Second, based on the orientation algorithm we have proposed, a novel localization algorithm is designed by decomposing the robot motion into uniform linear motion and uniform circular motion. The effectiveness of the proposed indoor localization method is verified via the real-world experiment using a tracked mobile robot developed in our laboratory.

Modern data center networks are usually constructed in multi-rooted tree topologies, which require the highly efficient multi-path load balancing to achieve high link utilization. Recent packet-level load balancer obtains high throughput by spraying packets to all paths, but it easily leads to the packet reordering under network asymmetry. The flow-level or flowlet-level load balancer avoids the packet reordering, while reducing the link utilization due to their inflexibility. To solve these problems, we design a Queueing Delay Aware Packet Spraying (QDAPS), that effectively mitigates the packet reordering for packet-level load balancer. QDAPS selects paths for packets according to the queueing delay of output buffer, and lets the packet arriving earlier be forwarded before the later packets to avoid packet reordering. We compare QDAPS with ECMP, LetFlow and RPS through NS2 simulation and Mininet implementation. The test results show that QDAPS reduces flow completion time (FCT) by ~30%-50% over the state-of-the-art load balancing mechanism.

Modern data-center applications generate a diverse mix of short and long flows with different performance requirements and weaknesses. The short flows are typically delay-sensitive but to suffer the head-of-line blocking and out-of-order problems. Recent solutions prioritize the short flows to meet their latency requirements, while damaging the throughput-sensitive long flows. To solve these problems, we design a Coding-based Adaptive Packet Spraying (CAPS) that effectively mitigates the negative impact of short and long flows on each other. To exploit the availability of multiple paths and avoid the head-of-line blocking, CAPS spreads the packets of short flows to all paths, while the long flows are limited to a few paths with Equal Cost Multi Path (ECMP). Meanwhile, to resolve the out-of-order problem with low overhead, CAPS encodes the short flows using forward error correction (FEC) technology and adjusts the coding redundancy according to the blocking probability. The coding layer is deployed between the TCP and IP layers, without any modifications on the existing TCP/IP protocols. The experimental results of NS2 simulation and Mininet implementation show that CAPS significantly reduces the average flow completion time of short flows by ~30% -70% over the state-of-the-art multipath transmission schemes and achieves the high throughput for long flows with negligible traffic overhead.

In order to mitigate the air pollution caused by traffic, the monitoring of on-road vehicle emission is really an urgent issue. The Vehicle Emission Remote Sensing System (VERSS) is a promising technology to solve this problem. But there is scarcely any available location strategy for traffic emission monitors yet to our knowledge, which restraints the use of monitors on a large scale of traffic network. In this paper, we make some efforts to solve a novel location problem in the transportation domain, that is, we look for the minimum subset of roads on which traffic emission monitors should be located, thus we can detect as many on-road vehicles as possible. We explicate how to transform the location problem to some graph problems and give the problem formulation mathematically. Then a two-step algorithm is designed to find the set of roads to locate monitors. The simulation test verify its availability. And in the last section some problems that should be studied further are presented at the end of the paper.

The real-time mapping of street atmospheric pollution concentration does play an important role because its knowledge is crucial for strategy-makers to make more effective control strategies to decrease urban atmospheric pollution and improving urban atmospheric environment. Combining the conventional methods (e.g. the dispersion model prediction and neural network prediction) and mobile measurement technology (e.g. the GMAP vehicle) which their characteristics are complementary, a linear model is proposed and then a fusion approach called weighting filter derived from the concept of Kalman filter. Moreover, a self-tuning regulator is introduced to adjust the parameters of filter for the changing noise statistical characteristics over time which mainly caused by season switch. The performances of asymptotic stability and asymptotic optimality are both mathematically proven. Finally a simulation test is conducted to verify this approach.