Alignment Model Research Articles

"Mission-Aligned Funds Flow": Effect on Clinical Departments.

Academic medical centers struggle with the high cost of care, reduced reimbursement, intense competition, and low profit margins. Many factors, including a high proportion of publicly insured patients, a model rewarding procedural specialties, and research and educational support burden, led to faculty salary inequities, physician disengagement, and difficulty recruiting. UC Davis Health implemented an aligned funds flow model in July 2021 to create a mission-aligned model in which all departments had financial margins to optimize recruitment, retention, research, and teaching. The 3-year experience (academic years 2021-2024) with this model at UC Davis Health was characterized by physician compensation, physician recruitment, and profit increases. Total collections for departments increased by 4% in the first year, 0.2% in the second year, and 11.3% in the third year of funds flow. Total productivity increased by 4.9% during the first year, 3.6% during the second year, and 8.4% during the third year. Salaries increased in all departmental categories in year 3. Productivity and collections per faculty member increased during the first year and were stable during the second and third years. Parity among procedural, primary care, and hospital-based service lines was improved because departmental revenue was agnostic to payer mix and hospital agreements were more formulaic. The hospital contribution to funds flow increased from $67 million in 2022 to $101 million in 2024. Regular communication and transparency are critical to ongoing trust and success with implementation of sustaining funds flow. The new model resulted in improved physician compensation and increased hiring. However, the implementation of funds flow had a negative fiscal effect on the academic medical center, and sustainability may require fine-tuning to balance affordability. The authors plan to convert outpatient primary care to productivity-based models and decrease time-limited support for new faculty from 2 years to 1 year.

Frameshift mutations in the mmpR5 gene can have a bedaquiline-susceptible phenotype by retaining a protein structure and function similar to wild-type Mycobacterium tuberculosis.

Bedaquiline (BDQ) is crucial for the treatment of rifampicin-resistant tuberculosis, yet resistance threatens its effectiveness, mainly linked to mutations in the mmpR5 (Rv0678) gene. While frameshift mutations are thought to produce non-functional proteins, we hypothesize that they can result in conserved proteins through late-stop codons or alternative reading frames and remain BDQ susceptible. We extracted 512 isolates harboring frameshift mutations in mmpR5 from the World Health Organization (WHO) catalog and 68 isolates with minimum inhibitory concentration (MIC) in mycobacterial growth indicator tube (MGIT) through a literature review. Using BioPython and AlphaFold2 we computed open (ORF) and alternative reading frames (ARFs) sequences and protein structures and assessed similarity to the wild type using an alignment and template modeling (TM)-score. Among the WHO 512 isolates, 24.8% were BDQ-sensitive. Out of 184 unique frameshift mutations with available nucleotide information, a late-stop codon in the ORF occurred for 32% of the mutations. Also, 40.7% resulted in a conserved sequence, through the ORF or one of the forward ARFs. In 68 isolates with available MGIT MIC data, the presence of late-stop codons in the ORF (OR 4.71, 95% CI 1.36-19.3) or a conserved reading frame (OR 10.4, 95% CI 2.07-102.9) were associated with BDQ sensitivity. Protein structures from the conserved sequences showed high similarity (TM > 0.8). We show that frameshift mutations may retain BDQ susceptibility through late-stop codons in the ORF or conserved ARFs. These findings could improve the prediction of the BDQ phenotype from genomic data and have important implications for treatment decisions. Research Foundation-Flanders, Academy of Medical Sciences, the Wellcome Trust, the Government Department of Business, Energy and Industrial Strategy, the British Heart Foundation and Diabetes UK, and the Global Challenges Research Fund.IMPORTANCETuberculosis (TB), caused by Mycobacterium tuberculosis, remains the deadliest infectious disease and is particularly challenging to treat when it becomes drug-resistant. Bedaquiline (BDQ) is a recently recommended core drug for treating drug-resistant TB. However, resistance to bedaquiline is already emerging, primarily due to mutations in the mmpR5 gene. Identifying which mutations cause resistance and which do not is a critical knowledge gap. In particular, little is known about the effect of frameshift mutations, typically thought to make TB bacteria resistant to bedaquiline by producing non-functional proteins. Yet, one-quarter of isolates with a frameshift mutation are still susceptible to bedaquiline. How the bacteria produce a functional protein despite the frameshift mutation is unknown. We analyzed over 500 frameshift mutations using computational methods to model their effects on protein structure and bedaquiline resistance. Our findings revealed that some frameshift mutations can still produce functional proteins, allowing bacteria to remain sensitive to bedaquiline. Specifically, bacteria can produce a functional protein despite frameshift mutations if the mutation occurs near the end of the protein or if an alternative reading frame is available. These insights improve our ability to interpret mutations associated with bedaquiline, the most important drug for drug-resistant TB, allowing more accurate and effective treatment decisions.

Third-order intrinsic alignment of SDSS BOSS LOWZ galaxies

Cosmic shear is a powerful probe of cosmology, but it is affected by the intrinsic alignment (IA) of galaxy shapes with the large-scale structure. Upcoming surveys such as Euclid and Vera C. Rubin Observatory's Legacy Survey of Space and Time (LSST) require an accurate understanding of IA, particularly for higher-order cosmic shear statistics that are vital for extracting the most cosmological information. In this paper, we report the first detection of third-order IA correlations using the LOWZ galaxy sample from the Sloan Digital Sky Survey (SDSS) Baryon Oscillation Spectroscopic Survey (BOSS). We compare our measurements with predictions from the MICE cosmological simulation and an analytical model inspired by the Non-linear Linear Alignment (NLA) model and informed by second-order correlations. We also explore the dependence of the third-order correlation on the galaxies' luminosity. We find that the amplitude $A_ IA $ of the IA signal is non-zero at the $4.7 level for scales between Mpc Mpc Mpc Mpc $ the inferred $A_ IA $ agrees both with the prediction from the simulation and estimates from second-order statistics within 1sigma but deviations arise at smaller scales. Our results demonstrate the feasibility of measuring third-order IA correlations and using them for constraining IA models. The agreement between second- and third-order IA constraints also opens the opportunity for a consistent joint analysis and IA self-calibration, promising tighter parameter constraints for upcoming cosmological surveys.

Campycins are novel broad-spectrum antibacterials killing Campylobacter jejuni

Pyocins are high molecular weight bacteriocins produced by Pseudomonas aeruginosa that can be retargeted to new bacterial species by exchanging the pyocin tail fibers with bacteriophage receptor binding proteins (RBPs). Here, we develop retargeted pyocins called campycins as new antibacterials to precisely and effectively kill the major foodborne pathogen Campylobacter jejuni. We used two diverse RBPs (H-fibers) encoded by CJIE1 prophages found in the genomes of C. jejuni strains CAMSA2147 and RM1221 to construct campycin 1 and campycin 2, respectively. Campycins 1 and 2 could target all C. jejuni strains tested due to complementary antibacterial spectra. In addition, both campycins led to more than 3 log reductions in C. jejuni counts under microaerobic conditions at 42 °C, whereas the killing efficiency was less efficient under anaerobic conditions at 5 °C. Furthermore, we discovered that both H-fibers used to construct the campycins bind to the essential major outer membrane protein (MOMP) present in all C. jejuni in a strain-specific manner. Protein sequence alignment and structural modeling suggest that the highly variable extracellular loops of MOMP form the binding sites of the diverse H-fibers. Further in silico analyses of 5000 MOMP sequences indicated that the protein falls into three major clades predicted to be targeted by either campycin 1 or campycin 2. Thus, campycins are promising antibacterials against C. jejuni and are expected to broadly target numerous strains of this human pathogen in nature and agriculture.Key points• Campycins are engineered R-type pyocins containing H-fibers from C. jejuni prophages• Campycins reduce C. jejuni counts by >3 logs at conditions promoting growth• Campycins bind to the essential outer membrane protein MOMP in a strain-dependent way

Adaptive multi-surrogate model for complex opto-mechanical system alignment

Owing to the inherent nonlinearity of surrogate models, they have been widely used in the alignment of complex opto-mechanical systems. Finding the most suitable surrogate model for misalignment wavefront errors is challenging. This study proposes an adaptive multi-surrogate model (AMSM) based on the concept of layer-by-layer approximation and adaptive selection and a compensation mechanism to facilitate the combination of the AMSM and non-dominated sorting genetic algorithm II (NSGA-II). For example, alignment simulations of an off-axis three-mirror anastigmatic (TMA) telescope and a catadioptric microscope objective were performed. The results show that the AMSM is well done in terms of approximate accuracy and efficiency. The application of the proposed method in the off-axis TMA telescope leads to 65.2% and 18.7% reductions in the full-field average RMS wavefront error compared with the traditional sensitivity table method (STM) and radial basis function (RBF) network surrogate model, respectively. Similarly, for the catadioptric microscope objective, there is a reduction of 43.5% and 22.4%, respectively. The AMSM will further promote the development of a surrogate model to align more complex and precise opto-mechanical systems.

Sinkhorn Distance Minimization for Adaptive Semi-Supervised Social Network Alignment.

Social network alignment, aiming at linking identical identities across different social platforms, is a fundamental task in social graph mining. Most existing approaches are supervised models and require a large number of manually labeled data, which are infeasible in practice considering the yawning gap between social platforms. Recently, isomorphism across social networks is incorporated as complementary to link identities from the distribution level, which contributes to alleviating the dependency on sample-level annotations. Adversarial learning is adopted to learn a shared projection function by minimizing the distance between two social distributions. However, the hypothesis of isomorphism might not always hold true as social user behaviors are generally unpredictable, and thus a shared projection function is insufficient to handle the sophisticated cross-platform correlations. In addition, adversarial learning suffers from training instability and uncertainty, which may hinder model performance. In this article, we propose a novel meta-learning-based social network alignment model Meta-SNA to effectively capture the isomorphism and the unique characteristics of each identity. Our motivation lies in learning a shared meta-model to preserve the global cross-platform knowledge and an adaptor to learn a specific projection function for each identity. Sinkhorn distance is further introduced as the distribution closeness measurement to tackle the limitations of adversarial learning, which owns an explicitly optimal solution and can be efficiently computed by the matrix scaling algorithm. Empirically, we evaluate the proposed model over multiple datasets, and the experimental results demonstrate the superiority of Meta-SNA.

Enhancing RNA-seq bias mitigation with the Gaussian self-benchmarking framework: towards unbiased sequencing data

BackgroundRNA sequencing is a vital technique for analyzing RNA behavior in cells, but it often suffers from various biases that distort the data. Traditional methods to address these biases are typically empirical and handle them individually, limiting their effectiveness. Our study introduces the Gaussian Self-Benchmarking (GSB) framework, a novel approach that leverages the natural distribution patterns of guanine (G) and cytosine (C) content in RNA to mitigate multiple biases simultaneously. This method is grounded in a theoretical model, organizing k-mers based on their GC content and applying a Gaussian model for alignment to ensure empirical sequencing data closely match their theoretical distribution.ResultsThe GSB framework demonstrated superior performance in mitigating sequencing biases compared to existing methods. Testing with synthetic RNA constructs and real human samples showed that the GSB approach not only addresses individual biases more effectively but also manages co-existing biases jointly. The framework’s reliance on accurately pre-determined parameters like mean and standard deviation of GC content distribution allows for a more precise representation of RNA samples. This results in improved accuracy and reliability of RNA sequencing data, enhancing our understanding of RNA behavior in health and disease.ConclusionsThe GSB framework presents a significant advancement in RNA sequencing analysis by providing a well-validated, multi-bias mitigation strategy. It functions independently from previously identified dataset flaws and sets a new standard for unbiased RNA sequencing results. This development enhances the reliability of RNA studies, broadening the potential for scientific breakthroughs in medicine and biology, particularly in genetic disease research and the development of targeted treatments.

Wfold: A new method for predicting RNA secondary structure with deep learning

Precise estimations of RNA secondary structures have the potential to reveal the various roles that non-coding RNAs play in regulating cellular activity. However, the mainstay of traditional RNA secondary structure prediction methods relies on thermos-dynamic models via free energy minimization, a laborious process that requires a lot of prior knowledge. Here, RNA secondary structure prediction using Wfold, an end-to-end deep learning-based approach, is suggested. Wfold is trained directly on annotated data and base-pairing criteria. It makes use of an image-like representation of RNA sequences, which an enhanced U-net incorporated with a transformer encoder can process effectively. Wfold eventually increases the accuracy of RNA secondary structure prediction by combining the benefits of self-attention mechanism's mining of long-range information with U-net's ability to gather local information. We compare Wfold's performance using RNA datasets that are within and across families. When trained and evaluated on different RNA families, it achieves a similar performance as the traditional methods, but dramatically outperforms the state-of-the-art methods on within-family datasets. Moreover, Wfold can also reliably forecast pseudoknots. The findings imply that Wfold may be useful for improving sequence alignment, functional annotations, and RNA structure modeling.

Research on Multi-Modal Music Score Alignment Model for Online Music Education

As music data storage becomes increasingly diverse in the era of big data, ensuring alignment of music works with the same semantics for online music education is crucial. To achieve this, a multi-modal music score alignment algorithm model based on deep learning was developed and optimized. Experimental results demonstrated that Note + DCO feature combination yielded the best MIDI input characteristics (mean value: 13.27 ms), whereas CQT feature comparison produced the best results for audio input (average: 12.85 ms). The ResNet-34 network was noted to have the most effective music score alignment effect with alignment errors averaging less than one frame. Compared with other algorithms, the proposed algorithm had the lowest average value of 9.28 ms, median value of 5.85 ms, and standard deviation of 20.17 ms. Actual music retrieval showed a Top-1 retrieval accuracy of 10.93% that was close to 11%. Overall, the proposed algorithm is significant for score alignment and music retrieval recognition in online music education.

Explanation of the problem of pulmonary embolism in relation to the optimization of modern algorithms for the actions of the family doctor

The purpose of the work is to ensure the improvement of the quality of education and educational services at a level that meets the expectations and needs of the family doctor, contributes to the optimization of the educational process on the problem of pulmonary embolism. Determination of clinical probability is an important component of management of patients with suspected PE. In clinical practice, the Canadian (P.S. Wells) and Geneva scales are most often used for this. For a patient with a high clinical probability of PE, multispiral computer tomography (MSCT) is of primary importance. With suspected high-risk PE, evidenced by shock or hypotension, CT angiography or bedside transthoracic echocardiography should be performed for diagnostic purposes. Today, magnetic resonance imaging is not recommended for the diagnosis of PE. Outpatient treatment should be carried out with PESI class I-II. Inpatient treatment - with PESI III-V class. Anticoagulant therapy (ACT), which should be started as early as possible at the stage of diagnosis, is the basis of VE treatment. Thrombolytic therapy should be carried out in the clinic of shock or in the presence of hemodynamic instability. As a basis for planning the educational process, it is desirable to take the constructive alignment model, which consists of the following three logically interconnected components. First of all, these are learning outcomes that must be aligned with the goals of the curriculum. Secondly, educational activities should be related to expected learning outcomes. An important component is assessment and feedback. The professional training of a general practitioner - a family doctor is aimed at acquiring new knowledge, deepening professional competences and improving practical skills to maintain an appropriate level of training for today's urgent problem - pulmonary embolism, taking into account the realities of wartime and the threat of repeated outbreaks of covid infection (CI)

The state as a model for AI control and alignment

The state as a model for AI control and alignment

Towards similar alignment and unique uniformity in collaborative filtering

Representation learning, with its desired properties, including Alignment and Uniformity, has recently emerged as an effective approach in collaborative filtering (CF) for recommender systems. Alignment measures the distance between positive pairs, and uniformity describes the distribution of all samples on the unit hypersphere. Despite the effectiveness of existing research, the challenges of insufficient alignment and uniformity bias persist in studies on the desired properties of recommender systems: (1) Sparse interaction information leads to an insufficient alignment representation. (2) Calculating the uniformity loss based on duplicate samples introduces bias, which affects the overall representation uniformity.Building on aforementioned challenges, we propose a Similar Alignment and Unique Uniformity model called SUAU. SUAU mitigates insufficient alignment by introducing additional item-similar item pairs for model training and employing a uniqueness strategy to prevent uniformity bias. More specifically, we propose a similar items generation method that identifies the most similar items based on the interaction behavior of the original items’ related users to form item-similar item pairs. The experimental results obtained on three highly sparse datasets demonstrate that the propose model achieves superior alignment and uniformity compared with state-of-the-art models, notably surpassing the best CF methods in terms of recommendation performance. We have open-sourced the source code on a publicly accessible website: https://github.com/ZzYUuuu/SUAU.

Dead-zone suppression method of NMOR atomic magnetometers based on alignment and orientation polarization

Nonlinear magneto-optical rotation (NMOR) atomic magnetometers demonstrate exceptional sensitivity in the geomagnetic environment, making them highly attractive for applications in resource exploration, biological research, and fundamental physics studies. Nevertheless, the presence of the “dead zone” hampers the magnetometer’s capacity to detect magnetic fields with sensitivity. In this paper, we present a method for effectively mitigating the “dead zone” by simultaneous detection of alignment and orientation polarization. Based on the standard formalism of density matrix and Liouville Equation, theoretical models of alignment and orientation resonance signals as a function of the magnetic field are developed. Additionally, due to the large light intensity used in the actual system, the alignment to orientation conversion (AOC) effect has been taken into account to reveal a more complete model. The influence of light intensity on the alignment and orientation signals are investigated. It is found that there is an optimal theoretical light intensity, which makes the suppression effect of the “dead zone” best. The theoretical model aligns well with the experimental phenomenon and successfully minimizes the extent of the “dead zone”.

Understanding the multi-wavelength thermal dust polarisation from the Orion molecular cloud in light of the radiative torque paradigm

Dust grains play a key role in various astrophysical processes and serve as indicators of interstellar medium structures, density, and mass. Understanding their physical properties and chemical composition is a crucial goal in astrophysics. Dust polarisation is a valuable tool for studying these properties. The radiative torque (RAT) paradigm, which includes radiative torque alignment (RAT-A) and radiative torque disruption (RAT-D), is essential to interpreting the dust polarisation data and constraining the fundamental properties of dust grains. However, it has been used primarily to interpret observations at a single wavelength. In this study, we analyse the thermal dust polarisation spectrum obtained from observations with SOFIA/HAWC+ and JCMT/POL-2 in the Orion molecular cloud 1 (OMC-1) region and compare the observational data with our numerical results using the RAT paradigm. In general, we show that the dense gas exhibits a positive spectral slope, whereas the warm regions show a negative one. We demonstrate that a one-layer dust (one-phase) model can only reproduce the observed spectra at certain locations and cannot match those with prominent V-shaped spectra (for which the degree of polarisation initially decreases with wavelength from 54 to ~300µm and then increases at longer wavelengths). To address this, we improved our model by incorporating two dust components (warm and cold) along the line of sight, resulting in a two-phase model. This improved model successfully reproduces the V-shaped spectra. The best model corresponds to a mixture composition of silicate and carbonaceous grains in the cold medium. Finally, by assuming the plausible model of grain alignment, we were able to infer the inclination angle of the magnetic fields in OMC-1. This approach is an important step towards a better understanding the physics of grain alignment and constraining 3D magnetic fields using dust polarisation spectra.

TriLA: Triple-Level Alignment Based Unsupervised Domain Adaptation for Joint Segmentation of Optic Disc and Optic Cup.

Cross-domain joint segmentation of optic disc and optic cup on fundus images is essential, yet challenging, for effective glaucoma screening. Although many unsupervised domain adaptation (UDA) methods have been proposed, these methods can hardly achieve complete domain alignment, leading to suboptimal performance. In this paper, we propose a triple-level alignment (TriLA) model to address this issue by aligning the source and target domains at the input level, feature level, and output level simultaneously. At the input level, a learnable Fourier domain adaptation (LFDA) module is developed to learn the cut-off frequency adaptively for frequency-domain translation. At the feature level, we disentangle the style and content features and align them in the corresponding feature spaces using consistency constraints. At the output level, we design a segmentation consistency constraint to emphasize the segmentation consistency across domains. The proposed model is trained on the RIGA+ dataset and widely evaluated on six different UDA scenarios. Our comprehensive results not only demonstrate that the proposed TriLA substantially outperforms other state-of-the-art UDA methods in joint segmentation of optic disc and optic cup, but also suggest the effectiveness of the triple-level alignment strategy.

Knowledge Graph Alignment Under Scarce Supervision: A General Framework With Active Cross-View Contrastive Learning.

Over recent years, a number of knowledge graphs (KGs) have emerged. Nevertheless, a KG can never reach full completeness. A viable approach to increase the coverage of a KG is KG alignment (KGA). The majority of previous efforts merely focus on the matching between entities, while largely neglect relations. Besides, they heavily rely on labeled data, which are difficult to obtain in practice. To address these issues, in this work, we put forward a general framework to simultaneously align entities and relations under scarce supervision. Our proposal consists of two main components, relation-enhanced active instance selection (RAS), and cross-view contrastive learning (CCL). RAS aims to select the most valuable instances to be labeled with the guidance of relations, while CCL contrasts cross-view representations to augment scarce supervision signals. Our proposal is agnostic to the underlying entity and relation alignment models, and can be used to improve their performance under limited supervision. We conduct experiments on a wide range of popular KG pairs, and the results demonstrate that our proposed model and its components can consistently boost the alignment performance under scarce supervision.

Data science and business value: Insight into the alignment process

We develop a comprehensive business - data science alignment model for deriving business value from data science initiatives, thus assisting business organizations in transforming digitally. With a purposive sample of 13 data science leaders in the capacities of C-level executives, semi-structured interviews were conducted to unfold their data science journey. The study extends the understanding from the business - IT/IS alignment literature to data science initiatives. The alignment dimensions and their components are extracted through the interview transcripts using a general inductive approach. We propose a business - data science alignment model as a dynamic and iterative process in which the data science initiatives are shaped while aligning with intellectual, operational, social, cultural, and societal dimensions to derive business value. The five most relevant components represent each of these alignment dimensions. The study emphasizes the groundwork to be done before investing in data science endeavors by capturing the learnings from the literature and practitioners. With the proposed data science alignment framework, our study offers to craft data science projects by guiding business executives in achieving business value.

Self-supervised Entity Alignment Model Based on Joint Embedding

Abstract Entity alignment is an important problem for constructing Web-scale KGs. Self-supervised entity alignment models represented by SelfKG, proposed in 2022, have eliminated the essential requirement of accurate alignment: human labeling, which significantly saves resources. However, when SelfKG scales up to improve entity alignment accuracy, it introduces neighbor noise, leading to a decrease in alignment performance. This indicates that the ability of self-supervised models to aggregate and utilize knowledge graph information needs further exploration. To address this issue, we proposed the following approach: using a Graph Convolutional Network(GCN) to collect global entity and structural information, and a Graph Attention Network(GAT) to collect local neighborhood information from subgraphs. Subsequently, we introduced a novel fusion method to merge the two parts through weighted sum and residual connection, and completed self-supervised entity alignment using the Relative Similarity Metric(RSM) mechanism proposed by SelfKG. This model has both the ability to scale up and the flexibility to avoid neighbor noise, resulting in a significant improvement in performance.

Modeling the Far-infrared Polarization Spectrum of a High-mass Star-forming Cloud

The polarization spectrum, or wavelength dependence of the polarization fraction, of interstellar dust emission provides important insights into the grain alignment mechanism of interstellar dust grains. We investigate the far-infrared polarization spectrum of a realistic simulated high-mass star-forming cloud under various models of grain alignment and emission. We find that neither a homogeneous grain alignment model nor a grain alignment model that includes collisional dealignment is able to produce the falling spectrum seen in observations. On the other hand, we find that a grain alignment model with grain alignment efficiency dependent on local temperature is capable of producing a falling spectrum that is in qualitative agreement with observations of OMC-1. For the model most in agreement with OMC-1, we find no correlation between the temperature and the slope of the polarization spectrum. However, we do find a positive correlation between the column density and the slope of the polarization spectrum. We suggest this latter correlation to be the result of wavelength-dependent polarization by absorption.

Time-series domain adaptation via sparse associative structure alignment: Learning invariance and variance

Domain adaptation on time-series data, which is often encountered in the field of industry, like anomaly detection and sensor data forecasting, but received limited attention in academia, is an important but challenging task in real-world scenarios. Most of the existing methods for time-series data use the covariate shift assumption for non-time-series data to extract the domain-invariant representation, but this assumption is hard to meet in practice due to the complex dependence among variables and a small change of the time lags may lead to a huge change of future values. To address this challenge, we leverage the stableness of causal structures among different domains. To further avoid the strong assumptions in causal discovery like linear non-Gaussian assumption, we relax it to mine the stable sparse associative structures instead of discovering the causal structures directly. Besides the domain-invariant structures, we also find that some domain-specific information like the strengths of the structures is important for prediction. Based on the aforementioned intuition, we extend the sparse associative structure alignment model in the conference version to the Sparse Associative Structure Alignment model with domain-specific information enhancement (SASA2 in short), which aligns the invariant unweighted spare associative structures and considers the variant information for time-series unsupervised domain adaptation. Specifically, we first generate the segment set to exclude the obstacle of offsets. Second, we extract the unweighted sparse associative structures via sparse attention mechanisms. Third, we extract the domain-specific information via an autoregressive module. Finally, we employ a unidirectional alignment restriction to guide the transformation from the source to the target. Moreover, we further provide a generalization analysis to show the theoretical superiority of our method. Compared with existing methods, our method yields state-of-the-art performance, with a 5% relative improvement in three real-world datasets, covering different applications: air quality, in-hospital healthcare, and anomaly detection. Furthermore, visualization results of sparse associative structures illustrate what knowledge can be transferred, boosting the transparency and interpretability of our method.