Search

In recent years, vision-language models have made significant strides, excelling in tasks like optical character recognition and geometric problem-solving. However, several critical issues remain: 1) Proprietary models often lack transparency about their architectures, while open-source models need more detailed ablations of their training strategies. 2) Pre-training data in open-source works is under-explored, with datasets added empirically, making the process cumbersome. 3) Fine-tuning often focuses on adding datasets, leading to diminishing returns. To address these issues, we propose the following contributions: 1) We trained a robust baseline model using the latest advancements in vision-language models, introducing effective improvements and conducting comprehensive ablation and validation for each technique. 2) Inspired by recent work on large language models, we filtered pre-training data using perplexity, selecting the lowest perplexity data for training. This approach allowed us to train on a curated 1M dataset, achieving competitive performance. 3) During visual instruction tuning, we used model soup on different datasets when adding more datasets yielded marginal improvements. These innovations resulted in a 9B parameter model that performs competitively with state-of-the-art models. Our strategies are efficient and lightweight, making them easily adoptable by the community., Comment: v1

Most large language models are fine-tuned using either expensive human-annotated data or GPT-4 generated data which cannot guarantee performance in certain domains. We argue that although the web-crawled data often has formatting errors causing semantic inaccuracies, it can still serve as a valuable source for high-quality supervised fine-tuning in specific domains without relying on advanced models like GPT-4. To this end, we create a paired training dataset automatically by aligning web-crawled data with a smaller set of high-quality data. By training a language model on this dataset, we can convert web data with irregular formats into high-quality ones. Our experiments show that training with the model-transformed data yields better results, surpassing training with only high-quality data by an average score of 9.4% in Chinese math problems. Additionally, our 7B model outperforms several open-source models larger than 32B and surpasses well-known closed-source models such as GPT-3.5, highlighting the efficacy of our approach.

For autonomous driving in highly dynamic environments, it is anticipated to predict the future behaviors of surrounding vehicles (SVs) and make safe and effective decisions. However, modeling the inherent coupling effect between the prediction and decision-making modules has been a long-standing challenge, especially when there is a need to maintain appropriate computational efficiency. To tackle these problems, we propose a novel integrated intention prediction and decision-making approach, which explicitly models the coupling relationship and achieves efficient computation. Specifically, a spectrum attention net is designed to predict the intentions of SVs by capturing the trends of each frequency component over time and their interrelations. Fast computation of the intention prediction module is attained as the predicted intentions are not decoded to trajectories in the executing process. Furthermore, the proximal policy optimization (PPO) algorithm is employed to address the non-stationary problem in the framework through a modest policy update enabled by a clipping mechanism within its objective function. On the basis of these developments, the intention prediction and decision-making modules are integrated through joint learning. Experiments are conducted in representative traffic scenarios, and the results reveal that the proposed integrated framework demonstrates superior performance over several deep reinforcement learning (DRL) baselines in terms of success rate, efficiency, and safety in driving tasks.

The explosion of massive urban data recently has provided us with a valuable opportunity to gain deeper insights into urban regions and the daily lives of residents. Urban region representation learning emerges as a crucial realm for fulfilling this task. Among deep learning approaches, graph neural networks (GNNs) have shown promise, given that city elements can be naturally represented as nodes with various connections between them as edges. However, many existing GNN approaches encounter challenges such as over-smoothing and limitations in capturing information from nodes in other regions, resulting in the loss of crucial urban information and a decline in region representation performance. To address these challenges, we leverage urban graph structure information and introduce a hierarchical graph pooling process called Coarsened Graph Attention Pooling (CGAP). CGAP features local attention units to create coarsened intermediate graphs and global features. Additionally, by incorporating urban region graphs and global features into a global attention layer, we harness relational information to enhance representation effectiveness. Furthermore, CGAP integrates region attributes such as Points of Interest (POIs) and inter-regional contexts like human mobility, enabling the exploitation of multi-modal urban data for more comprehensive representation learning. Experiments on three downstream tasks related to the UN Sustainable Development Goals validate the effectiveness of region representations learned by our approach. Experimental results and analyses demonstrate that CGAP excels in various socioeconomic prediction tasks compared to competitive baselines.

Automatic vessel segmentation is paramount for developing next-generation interventional navigation systems. However, current approaches suffer from suboptimal segmentation performances due to significant challenges in intraoperative images (i.e., low signal-to-noise ratio, small or slender vessels, and strong interference). In this paper, a novel spatial-frequency learning and topological channel interaction network (SPIRONet) is proposed to address the above issues. Specifically, dual encoders are utilized to comprehensively capture local spatial and global frequency vessel features. Then, a cross-attention fusion module is introduced to effectively fuse spatial and frequency features, thereby enhancing feature discriminability. Furthermore, a topological channel interaction module is designed to filter out task-irrelevant responses based on graph neural networks. Extensive experimental results on several challenging datasets (CADSA, CAXF, DCA1, and XCAD) demonstrate state-of-the-art performances of our method. Moreover, the inference speed of SPIRONet is 21 FPS with a 512x512 input size, surpassing clinical real-time requirements (6~12FPS). These promising outcomes indicate SPIRONet's potential for integration into vascular interventional navigation systems. Code is available at https://github.com/Dxhuang-CASIA/SPIRONet.

Ratings of a user to most items in recommender systems are usually missing not at random (MNAR), largely because users are free to choose which items to rate. To achieve unbiased learning of the prediction model under MNAR data, three typical solutions have been proposed, including error-imputation-based (EIB), inverse-propensity-scoring (IPS), and doubly robust (DR) methods. However, these methods ignore an alternative form of bias caused by the inconsistency between the observed ratings and the users' true preferences, also known as noisy feedback or outcome measurement errors (OME), e.g., due to public opinion or low-quality data collection process. In this work, we study intersectional threats to the unbiased learning of the prediction model from data MNAR and OME in the collected data. First, we design OME-EIB, OME-IPS, and OME-DR estimators, which largely extend the existing estimators to combat OME in real-world recommendation scenarios. Next, we theoretically prove the unbiasedness and generalization bound of the proposed estimators. We further propose an alternate denoising training approach to achieve unbiased learning of the prediction model under MNAR data with OME. Extensive experiments are conducted on three real-world datasets and one semi-synthetic dataset to show the effectiveness of our proposed approaches. The code is available at https://github.com/haoxuanli-pku/KDD24-OME-DR., Comment: KDD 24 Research Track Paper

Predicting socioeconomic indicators within urban regions is crucial for fostering inclusivity, resilience, and sustainability in cities and human settlements. While pioneering studies have attempted to leverage multi-modal data for socioeconomic prediction, jointly exploring their underlying semantics remains a significant challenge. To address the gap, this paper introduces a Multi-Semantic Contrastive Learning (MuseCL) framework for fine-grained urban region profiling and socioeconomic prediction. Within this framework, we initiate the process by constructing contrastive sample pairs for street view and remote sensing images, capitalizing on the similarities in human mobility and Point of Interest (POI) distribution to derive semantic features from the visual modality. Additionally, we extract semantic insights from POI texts embedded within these regions, employing a pre-trained text encoder. To merge the acquired visual and textual features, we devise an innovative cross-modality-based attentional fusion module, which leverages a contrastive mechanism for integration. Experimental results across multiple cities and indicators consistently highlight the superiority of MuseCL, demonstrating an average improvement of 10% in $R^2$ compared to various competitive baseline models. The code of this work is publicly available at https://github.com/XixianYong/MuseCL.

The Bell nonlocality and entanglement are two kinds of quantum correlations in quantum systems. Due to the recent upgrade in Beijing Spectrometer III (BESIII) experiment, it is possible to explore the nonlocality and entanglement in hyperon-antihyperon systems produced in electron-positron annihilation with high precision data. We provide a systematic method for studying quantum correlations in spin-1/2 hyperon-antihyperon systems through the measures for the nonlocality and entanglement. We find that with nonvanishing polarizations of the hyperon and its antihyperon, the kinematic region of nonlocality in the hyperon-antihyperon system is more restricted than the $\tau^{+}\tau^{-}$ system in which polarizations of $\tau$ leptons are vanishing. We also present an experimental proposal to probe the nonlocality and entanglement in hyperon-antihyperon systems at BSEIII., Comment: 9 pages, 4 figures, 4 tables. We corrected a few errors in plotting figures from analytical formula. Some results in tables read from figures have also been corrected. A new table (Table III) was added for the maximum concurrence and their corresponding angles. A few references were added

When evaluating the effectiveness of a drug, a Randomized Controlled Trial (RCT) is often considered the gold standard due to its perfect randomization. While RCT assures strong internal validity, its restricted external validity poses challenges in extending treatment effects to the broader real-world population due to possible heterogeneity in covariates. In this paper, we introduce a procedure to generalize the RCT findings to the real-world trial-eligible population based on the adaption of existing statistical methods. We utilized the augmented inversed probability of sampling weighting (AIPSW) estimator for the estimation and omitted variable bias framework to assess the robustness of the estimate against the assumption violation caused by potentially unmeasured confounders. We analyzed an RCT comparing the effectiveness of lowering hypertension between Songling Xuemaikang Capsule (SXC), a traditional Chinese medicine (TCM), and Losartan as an illustration. The generalization results indicated that although SXC is less effective in lowering blood pressure than Losartan on week 2, week 4, and week 6, there is no statistically significant difference among the trial-eligible population at week 8, and the generalization is robust against potential unmeasured confounders., Comment: 21 pages, 3 figures, 3tables

Publication bias (PB) poses a significant threat to meta-analysis, as studies yielding notable results are more likely to be published in scientific journals. Sensitivity analysis provides a flexible method to address PB and to examine the impact of unpublished studies. A selection model based on t-statistics to sensitivity analysis is proposed by Copas. This t-statistics selection model is interpretable and enables the modeling of biased publication sampling across studies, as indicated by the asymmetry in the funnel-plot. In meta-analysis of diagnostic studies, the summary receiver operating characteristic curve is an essential tool for synthesizing the bivariate outcomes of sensitivity and specificity reported by individual studies. Previous studies address PB upon the bivariate normal model but these methods rely on the normal approximation for the empirical logit-transformed sensitivity and specificity, which is not suitable for sparse data scenarios. Compared to the bivariate normal model, the bivariate binomial model which replaces the normal approximation in the within-study model with the exact within-study model has better finite sample properties. In this study, we applied the Copas t-statistics selection model to the meta-analysis of diagnostic studies using the bivariate binomial model. To our knowledge, this is the first study to apply the Copas t-statistics selection model to the bivariate binomial model. We have evaluated our proposed method through several real-world meta-analyses of diagnostic studies and simulation studies.

Instruction Fine-Tuning enhances pre-trained language models from basic next-word prediction to complex instruction-following. However, existing One-off Instruction Fine-Tuning (One-off IFT) method, applied on a diverse instruction, may not effectively boost models' adherence to instructions due to the simultaneous handling of varying instruction complexities. To improve this, Phased Instruction Fine-Tuning (Phased IFT) is proposed, based on the idea that learning to follow instructions is a gradual process. It assesses instruction difficulty using GPT-4, divides the instruction data into subsets of increasing difficulty, and uptrains the model sequentially on these subsets. Experiments with Llama-2 7B/13B/70B, Llama3 8/70B and Mistral-7B models using Alpaca data show that Phased IFT significantly outperforms One-off IFT, supporting the progressive alignment hypothesis and providing a simple and efficient way to enhance large language models. Codes and datasets from our experiments are freely available at https://github.com/xubuvd/PhasedSFT., Comment: The final version, to be appear at ACL 2024 Findings

Layered van der Waals transition metal dichalcogenides (TMDCs), generally composed of three atomic X-M-X planes in each layer (M = transition metal, X = chalcogen), provide versatile platforms for exploring diverse quantum phenomena. In each MX2 layer, the M-X bonds are predominantly covalent in nature, as a result, the cleavage of TMDC crystals always occurring between the layers. Here we report the controllable realization of fractional-layer WTe2 via an in-situ scanning tunnelling microscopy (STM) tip manipulation technique. By applying STM tip pulses, hundreds of the topmost Te atoms are removed to form a nanoscale monolayer Te pit in the 1T'-WTe2, thus realizing a brand-new 2/3-layer WTe2. Such a unique configuration undergoes a spontaneous atomic reconstruction, yielding an energy-dependent unidirectional charge-density-wave state with the wavevector and geometry quite distinct from that of pristine 1T'-WTe2. Our results expand the conventional understanding of the TMDCs and are expected to stimulate the research on extraordinary structures and properties based on fractional-layer TMDCs., Comment: 4 figures in main text

Current text-to-image diffusion models have achieved groundbreaking results in image generation tasks. However, the unavoidable inclusion of sensitive information during pre-training introduces significant risks such as copyright infringement and privacy violations in the generated images. Machine Unlearning (MU) provides a effective way to the sensitive concepts captured by the model, has been shown to be a promising approach to addressing these issues. Nonetheless, existing MU methods for concept erasure encounter two primary bottlenecks: 1) generalization issues, where concept erasure is effective only for the data within the unlearn set, and prompts outside the unlearn set often still result in the generation of sensitive concepts; and 2) utility drop, where erasing target concepts significantly degrades the model's performance. To this end, this paper first proposes a concept domain correction framework for unlearning concepts in diffusion models. By aligning the output domains of sensitive concepts and anchor concepts through adversarial training, we enhance the generalizability of the unlearning results. Secondly, we devise a concept-preserving scheme based on gradient surgery. This approach alleviates the parts of the unlearning gradient that contradict the relearning gradient, ensuring that the process of unlearning minimally disrupts the model's performance. Finally, extensive experiments validate the effectiveness of our model, demonstrating our method's capability to address the challenges of concept unlearning in diffusion models while preserving model utility.

Recent advancements in large vision-language models (LVLMs), such as GPT4-V and LLaVA, have been substantial. LLaVA's modular architecture, in particular, offers a blend of simplicity and efficiency. Recent works mainly focus on introducing more pre-training and instruction tuning data to improve model's performance. This paper delves into the often-neglected aspects of data efficiency during pre-training and the selection process for instruction tuning datasets. Our research indicates that merely increasing the size of pre-training data does not guarantee improved performance and may, in fact, lead to its degradation. Furthermore, we have established a pipeline to pinpoint the most efficient instruction tuning (SFT) dataset, implying that not all SFT data utilized in existing studies are necessary. The primary objective of this paper is not to introduce a state-of-the-art model, but rather to serve as a roadmap for future research, aiming to optimize data usage during pre-training and fine-tuning processes to enhance the performance of vision-language models.

Publication bias (PB) is one of the serious issues in meta-analysis. Many existing methods dealing with PB are based on the normal-normal (NN) random-effects model assuming normal models in both the within-study and the between-study levels. For rare-event meta-analysis where the data contain rare occurrences of event, the standard NN random-effects model may perform poorly. Instead, the generalized linear mixed effects model (GLMM) using the exact within-study model is recommended. However, no method has been proposed for dealing with PB in rare-event meta-analysis using the GLMM. In this paper, we propose sensitivity analysis methods for evaluating the impact of PB on the GLMM based on the famous Copas-Heckman-type selection model. The proposed methods can be easily implemented with the standard software coring the nonlinear mixed-effects model. We use a real-world example to show how the usefulness of the proposed methods in evaluating the potential impact of PB in meta-analysis of the log-transformed odds ratio based on the GLMM using the non-central hypergeometric or binomial distribution as the within-study model. An extension of the proposed method is also introduced for evaluating PB in meta-analysis of proportion based on the GLMM with the binomial within-study model.

Accurate recognition of human motion intention (HMI) is beneficial for exoskeleton robots to improve the wearing comfort level and achieve natural human-robot interaction. A classifier trained on labeled source subjects (domains) performs poorly on unlabeled target subject since the difference in individual motor characteristics. The unsupervised domain adaptation (UDA) method has become an effective way to this problem. However, the labeled data are collected from multiple source subjects that might be different not only from the target subject but also from each other. The current UDA methods for HMI recognition ignore the difference between each source subject, which reduces the classification accuracy. Therefore, this paper considers the differences between source subjects and develops a novel theory and algorithm for UDA to recognize HMI, where the margin disparity discrepancy (MDD) is extended to multi-source UDA theory and a novel weight-aware-based multi-source UDA algorithm (WMDD) is proposed. The source domain weight, which can be adjusted adaptively by the MDD between each source subject and target subject, is incorporated into UDA to measure the differences between source subjects. The developed multi-source UDA theory is theoretical and the generalization error on target subject is guaranteed. The theory can be transformed into an optimization problem for UDA, successfully bridging the gap between theory and algorithm. Moreover, a lightweight network is employed to guarantee the real-time of classification and the adversarial learning between feature generator and ensemble classifiers is utilized to further improve the generalization ability. The extensive experiments verify theoretical analysis and show that WMDD outperforms previous UDA methods on HMI recognition tasks., Comment: 13 pages, 5 figures

In this paper, we consider the problem of visual representation learning for computational pathology, by exploiting large-scale image-text pairs gathered from public resources, along with the domain-specific knowledge in pathology. Specifically, we make the following contributions: (i) We curate a pathology knowledge tree that consists of 50,470 informative attributes for 4,718 diseases requiring pathology diagnosis from 32 human tissues. To our knowledge, this is the first comprehensive structured pathology knowledge base; (ii) We develop a knowledge-enhanced visual-language pretraining approach, where we first project pathology-specific knowledge into latent embedding space via a language model, and use it to guide the visual representation learning; (iii) We conduct thorough experiments to validate the effectiveness of our proposed components, demonstrating significant performance improvement on various downstream tasks, including cross-modal retrieval, zero-shot classification on pathology patches, and zero-shot tumor subtyping on whole slide images (WSIs)., Comment: ECCV2024(Oral)

Natural Language Processing (NLP) is witnessing a remarkable breakthrough driven by the success of Large Language Models (LLMs). LLMs have gained significant attention across academia and industry for their versatile applications in text generation, question answering, and text summarization. As the landscape of NLP evolves with an increasing number of domain-specific LLMs employing diverse techniques and trained on various corpus, evaluating performance of these models becomes paramount. To quantify the performance, it's crucial to have a comprehensive grasp of existing metrics. Among the evaluation, metrics which quantifying the performance of LLMs play a pivotal role. This paper offers a comprehensive exploration of LLM evaluation from a metrics perspective, providing insights into the selection and interpretation of metrics currently in use. Our main goal is to elucidate their mathematical formulations and statistical interpretations. We shed light on the application of these metrics using recent Biomedical LLMs. Additionally, we offer a succinct comparison of these metrics, aiding researchers in selecting appropriate metrics for diverse tasks. The overarching goal is to furnish researchers with a pragmatic guide for effective LLM evaluation and metric selection, thereby advancing the understanding and application of these large language models.

Phase singularities are phase-indeterminate points where wave amplitudes are zero, which manifest as phase vertices or wavefront dislocations. In the realm of optical and electron beams, the phase singularity has been extensively explored, demonstrating a profound connection to orbital angular momentum. Direct local imaging of the impact of orbital angular momentum on phase singularities at the nanoscale, however, remains a challenge and has yet to be achieved. Here, we study the role of orbital angular momentum in phase singularities in graphene, particularly at the atomic level, through scanning tunneling microscopy and spectroscopy. Our experiments demonstrate that the scatterings between different orbital angular momentum states, which are induced by local rotational symmetry-breaking potentials, can generate additional phase singularity, and result in robust single wavefront dislocation in real space. Our results pave the way for exploring the effects of orbital degree of freedom on quantum phases in quasiparticle interference processes., Comment: 28 pages, 3 figures, 10 extended figures

In this work, we present a reward-driven automated curriculum reinforcement learning approach for interaction-aware self-driving at unsignalized intersections, taking into account the uncertainties associated with surrounding vehicles (SVs). These uncertainties encompass the uncertainty of SVs' driving intention and also the quantity of SVs. To deal with this problem, the curriculum set is specifically designed to accommodate a progressively increasing number of SVs. By implementing an automated curriculum selection mechanism, the importance weights are rationally allocated across various curricula, thereby facilitating improved sample efficiency and training outcomes. Furthermore, the reward function is meticulously designed to guide the agent towards effective policy exploration. Thus the proposed framework could proactively address the above uncertainties at unsignalized intersections by employing the automated curriculum learning technique that progressively increases task difficulty, and this ensures safe self-driving through effective interaction with SVs. Comparative experiments are conducted in $Highway\_Env$, and the results indicate that our approach achieves the highest task success rate, attains strong robustness to initialization parameters of the curriculum selection module, and exhibits superior adaptability to diverse situational configurations at unsignalized intersections. Furthermore, the effectiveness of the proposed method is validated using the high-fidelity CARLA simulator., Comment: 8 pages, 6 figures

Continuously and widely tunable lasers actively stabilized on a frequency reference are broadly employed in atomic, molecular and optical (AMO) physics. The frequency-stabilized optical frequency comb (OFC) provides a novel optical frequency reference with a broadband spectrum that meets the requirement of laser frequency stabilization. Therefore, we demonstrate a frequency-stabilized and precisely tunable laser system based on it. In this scheme, the laser frequency locked to the OFC is driven to jump over the ambiguity zones, which blocks the wide tuning of the locked laser, and tuned until the mode hopping happens with the always-activated feedback loop. Meanwhile, we compensate the gap of the frequency jump with a synchronized acoustic optical modulator to ensure the continuity. This scheme is applied to an external cavity diode laser (ECDL) and we achieve tuning at a rate of about 7 GHz/s with some readily available commercial electronics. Furthermore, we tune the frequency-stabilized laser only with the feedback of diode current and its average tuning speed can exceed 100 GHz/s. Due to the resource-efficient configuration and the simplicity of completion, this scheme can be referenced and find wide applications in AMO experiments.

The generation of cold molecules is a core topic in the field of cold atoms and molecules, which has advanced relevant research like ultracold chemistry, quantum computation, and quantum metrology. With high atomic phase space density, optical dipole trap has been widely performed to prepare and trap cold molecules, and can also be further developed for multiple cold molecule formation and dynamics study. In this work, Rb2 molecules are photoassociated in the magneto-optical trap to obtain precise rovibrational spectroscopy, which provides accurate numerical references for multiple photoassociations. By achieving the harsh requirements of photoassociation in the optical dipole trap, the cold molecule photoassociation process is well explored, and different rovibrational cold molecules are formed in the optical dipole trap for the first time. This method can be universally extended to simultaneously photoassociate various molecules with different internal states or atomic species in just one optical dipole trap, and then advance generous cold molecule research such as cold molecule collision dynamics., Comment: 6 pages, 5 figures

In randomized controlled trials (RCT) with time-to-event outcomes, intercurrent events occur as semi-competing/competing events, and they could affect the hazard of outcomes or render outcomes ill-defined. Although five strategies have been proposed in ICH E9 (R1) addendum to address intercurrent events in RCT, they did not readily extend to the context of time-to-event data for studying causal effects. In this study, we show how to define, estimate, and infer the time-dependent cumulative incidence of outcome events in such contexts for obtaining causal interpretations. Specifically, we derive the mathematical forms of the scientific objective (i.e., causal estimands) under the five strategies and clarify the required data structure to identify these causal estimands. Furthermore, we summarize estimation and inference methods for these causal estimands by adopting methodologies in survival analysis, including analytic formulas for asymptotic analysis and hypothesis testing. We illustrate our methods with the LEADER Trial on investigating the effect of liraglutide on cardiovascular outcomes. Studies of multiple endpoints and combining strategies to address multiple intercurrent events can help practitioners understand treatment effects more comprehensively.

Medical image segmentation takes an important position in various clinical applications. Deep learning has emerged as the predominant solution for automated segmentation of volumetric medical images. 2.5D-based segmentation models bridge computational efficiency of 2D-based models and spatial perception capabilities of 3D-based models. However, prevailing 2.5D-based models often treat each slice equally, failing to effectively learn and exploit inter-slice information, resulting in suboptimal segmentation performances. In this paper, a novel Momentum encoder-based inter-slice fusion transformer (MOSformer) is proposed to overcome this issue by leveraging inter-slice information at multi-scale feature maps extracted by different encoders. Specifically, dual encoders are employed to enhance feature distinguishability among different slices. One of the encoders is moving-averaged to maintain the consistency of slice representations. Moreover, an IF-Swin transformer module is developed to fuse inter-slice multi-scale features. The MOSformer is evaluated on three benchmark datasets (Synapse, ACDC, and AMOS), establishing a new state-of-the-art with 85.63%, 92.19%, and 85.43% of DSC, respectively. These promising results indicate its competitiveness in medical image segmentation. Codes and models of MOSformer will be made publicly available upon acceptance., Comment: Under Review

Retrieval models aim at selecting a small set of item candidates which match the preference of a given user. They play a vital role in large-scale recommender systems since subsequent models such as rankers highly depend on the quality of item candidates. However, most existing retrieval models employ a single-round inference paradigm, which may not adequately capture the dynamic nature of user preferences and stuck in one area in the item space. In this paper, we propose Ada-Retrieval, an adaptive multi-round retrieval paradigm for recommender systems that iteratively refines user representations to better capture potential candidates in the full item space. Ada-Retrieval comprises two key modules: the item representation adapter and the user representation adapter, designed to inject context information into items' and users' representations. The framework maintains a model-agnostic design, allowing seamless integration with various backbone models such as RNNs or Transformers. We perform experiments on three widely used public datasets, incorporating five powerful sequential recommenders as backbone models. Our results demonstrate that Ada-Retrieval significantly enhances the performance of various base models, with consistent improvements observed across different datasets. Our code and data are publicly available at: https://github.com/ll0ruc/Ada-Retrieval., Comment: 9 pages, Accepted to AAAI2024

Artificial light plays an integral role in modern cities, significantly enhancing human productivity and the efficiency of civilization. However, excessive illumination can lead to light pollution, posing non-negligible threats to economic burdens, ecosystems, and human health. Despite its critical importance, the exploration of its causes remains relatively limited within the field of artificial intelligence, leaving an incomplete understanding of the factors contributing to light pollution and sustainable illumination planning distant. To address this gap, we introduce a novel framework named Causally Aware Generative Adversarial Networks (CAGAN). This innovative approach aims to uncover the fundamental drivers of light pollution within cities and offer intelligent solutions for optimal illumination resource allocation in the context of sustainable urban development. We commence by examining light pollution across 33,593 residential areas in seven global metropolises. Our findings reveal substantial influences on light pollution levels from various building types, notably grasslands, commercial centers and residential buildings as significant contributors. These discovered causal relationships are seamlessly integrated into the generative modeling framework, guiding the process of generating light pollution maps for diverse residential areas. Extensive experiments showcase CAGAN's potential to inform and guide the implementation of effective strategies to mitigate light pollution. Our code and data are publicly available at https://github.com/zhangyuuao/Light_Pollution_CAGAN., Comment: 9pages, 9figures, accepted by AAAI2024, AI for Social Impact (Special Track)

In a molecule formed by two atoms, energy difference between bonding and antibonding orbitals should depend on distance of the two atoms. However, exploring molecular orbitals of two natural atoms with tunable distance has remained an outstanding experimental challenge. Graphene quantum dots (GQDs) can be viewed as relativistic artificial atoms, therefore, offering a unique platform to study molecular physics. Here, through scanning tunneling microscope (STM), we create and directly visualize the formation process of relativistic artificial molecules based on two coupled GQDs with tunable distance. Our study indicates that energy difference between the bonding and antibonding orbitals of the lowest quasibound state increases linearly with inverse distance of the two GQDs due to the relativistic nature of the artificial molecule. For quasibound states with higher orbital momenta, the coupling between these states leads to half-energy spacing of the confined states because the length of the molecular-like orbit is about twice that of the atomic-like orbit. Evolution from ring-like whispering-gallery modes in the artificial atoms to figure-eight orbitals in the artificial molecules is directly imaged. The ability to resolve the coupling and orbitals of the relativistic artificial molecule at the nanoscale level yields insights into the behavior of quantum-relativistic matter.

Accurately modeling detailed interactions between human/hand and object is an appealing yet challenging task. Current multi-view capture systems are only capable of reconstructing multiple subjects into a single, unified mesh, which fails to model the states of each instance individually during interactions. To address this, previous methods use template-based representations to track human/hand and object. However, the quality of the reconstructions is limited by the descriptive capabilities of the templates so that these methods are inherently struggle with geometry details, pressing deformations and invisible contact surfaces. In this work, we propose an end-to-end Instance-aware Human-Object Interactions recovery (Ins-HOI) framework by introducing an instance-level occupancy field representation. However, the real-captured data is presented as a holistic mesh, unable to provide instance-level supervision. To address this, we further propose a complementary training strategy that leverages synthetic data to introduce instance-level shape priors, enabling the disentanglement of occupancy fields for different instances. Specifically, synthetic data, created by randomly combining individual scans of humans/hands and objects, guides the network to learn a coarse prior of instances. Meanwhile, real-captured data helps in learning the overall geometry and restricting interpenetration in contact areas. As demonstrated in experiments, our method Ins-HOI supports instance-level reconstruction and provides reasonable and realistic invisible contact surfaces even in cases of extremely close interaction. To facilitate the research of this task, we collect a large-scale, high-fidelity 3D scan dataset, including 5.2k high-quality scans with real-world human-chair and hand-object interactions. The code and data will be public for research purposes., Comment: Project Page: https://jiajunzhang16.github.io/ins-hoi/ , Code and Dataset Page: https://github.com/jiajunzhang16/ins-hoi

The problem of packing as many subgraphs isomorphic to $H \in \mathcal H$ as possible in a graph for a class $\mathcal H$ of graphs is well studied in the literature. Both vertex-disjoint and edge-disjoint versions are known to be NP-complete for $H$ that contains at least three vertices and at least three edges, respectively. In this paper, we consider ``list variants'' of these problems: Given a graph $G$, an integer $k$, and a collection $\mathcal L_{\mathcal H}$ of subgraphs of $G$ isomorphic to some $H \in \mathcal H$, the goal is to compute $k$ subgraphs in $\mathcal L_{\mathcal H}$ that are pairwise vertex- or edge-disjoint. We show several positive and negative results, focusing on classes of sparse graphs, such as bounded-degree graphs, planar graphs, and bounded-treewidth graphs.

Offline reinforcement learning (RL) faces a significant challenge of distribution shift. Model-free offline RL penalizes the Q value for out-of-distribution (OOD) data or constrains the policy closed to the behavior policy to tackle this problem, but this inhibits the exploration of the OOD region. Model-based offline RL, which uses the trained environment model to generate more OOD data and performs conservative policy optimization within that model, has become an effective method for this problem. However, the current model-based algorithms rarely consider agent robustness when incorporating conservatism into policy. Therefore, the new model-based offline algorithm with a conservative Bellman operator (MICRO) is proposed. This method trades off performance and robustness via introducing the robust Bellman operator into the algorithm. Compared with previous model-based algorithms with robust adversarial models, MICRO can significantly reduce the computation cost by only choosing the minimal Q value in the state uncertainty set. Extensive experiments demonstrate that MICRO outperforms prior RL algorithms in offline RL benchmark and is considerably robust to adversarial perturbations., Comment: Accepted by IJCAI 2024 (the 33rd International Joint Conference on Artificial Intelligence)