Search

The use of four-day school weeks (4dsw) in the United States has expanded rapidly over the past two decades. Previous work examines the impact of 4dsw on student outcomes, but little research to date examines the effect on school employees even though schools in some locales have adopted 4dsw to recruit and retain staff. This paper examines the effect of 4dsw adoption in Oregon, a state with widespread 4dsw use, on teacher and other school staff retention by leveraging a staggered roll-out of the schedule using a difference-in-differences design. We find that adopting a four-day week increased turnover among teachers, but that turnover among non-teaching staff was largely unaffected. The findings suggest that policymakers interested in implementing 4dsw for improved school employee retention should exercise caution and be attentive to the full set of incentives offered to staff.

This study provides the first large-scale quantitative exploration of mathematical language use in upper elementary U.S. classrooms. Our approach employs natural language processing techniques to describe variation in teachers' and students' use of mathematical language in 1,657 fourth and fifth grade lessons in 317 classrooms in four districts over three years. Students' exposure to mathematical language varies substantially across lessons and between teachers. Results suggest that teacher modeling, defined as the density of mathematical terms in teacher talk, does not substantially cause students to uptake mathematical language, but that teachers may encourage student use of mathematical vocabulary by means other than mere modeling or exposure. However, we also find that teachers who use more mathematical language are more effective at raising student test scores. These findings reveal that teachers who use more mathematical vocabulary are more effective math teachers.

Dual enrollment (DE) is one of the fastest growing programs that support the high school-to-college transition. Yet, there is limited empirical evidence about its impact on either students' college application choices or admission outcomes. Using a fuzzy regression discontinuity approach on data from two cohorts of ninth-grade students in one anonymous state, we found that taking DE credits increased the likelihood of applying to highly selective in-state four-year institutions. Attempting DE credits also increased the probability of gaining admission to a highly selective in-state four-year college. Heterogeneous analysis further indicates that the gains were extended across Black, Latinx, and white student populations.

As one of the automotive sensors that have emerged in recent years, 4D millimeter-wave radar has a higher resolution than conventional 3D radar and provides precise elevation measurements. But its point clouds are still sparse and noisy, making it challenging to meet the requirements of autonomous driving. Camera, as another commonly used sensor, can capture rich semantic information. As a result, the fusion of 4D radar and camera can provide an affordable and robust perception solution for autonomous driving systems. However, previous radar-camera fusion methods have not yet been thoroughly investigated, resulting in a large performance gap compared to LiDAR-based methods. Specifically, they ignore the feature-blurring problem and do not deeply interact with image semantic information. To this end, we present a simple but effective multi-stage sampling fusion (MSSF) network based on 4D radar and camera. On the one hand, we design a fusion block that can deeply interact point cloud features with image features, and can be applied to commonly used single-modal backbones in a plug-and-play manner. The fusion block encompasses two types, namely, simple feature fusion (SFF) and multiscale deformable feature fusion (MSDFF). The SFF is easy to implement, while the MSDFF has stronger fusion abilities. On the other hand, we propose a semantic-guided head to perform foreground-background segmentation on voxels with voxel feature re-weighting, further alleviating the problem of feature blurring. Extensive experiments on the View-of-Delft (VoD) and TJ4DRadset datasets demonstrate the effectiveness of our MSSF. Notably, compared to state-of-the-art methods, MSSF achieves a 7.0% and 4.0% improvement in 3D mean average precision on the VoD and TJ4DRadSet datasets, respectively. It even surpasses classical LiDAR-based methods on the VoD dataset.

Human motion generation plays a vital role in applications such as digital humans and humanoid robot control. However, most existing approaches disregard physics constraints, leading to the frequent production of physically implausible motions with pronounced artifacts such as floating and foot sliding. In this paper, we propose \textbf{Morph}, a \textbf{Mo}tion-f\textbf{r}ee \textbf{ph}ysics optimization framework, comprising a Motion Generator and a Motion Physics Refinement module, for enhancing physical plausibility without relying on costly real-world motion data. Specifically, the Motion Generator is responsible for providing large-scale synthetic motion data, while the Motion Physics Refinement Module utilizes these synthetic data to train a motion imitator within a physics simulator, enforcing physical constraints to project the noisy motions into a physically-plausible space. These physically refined motions, in turn, are used to fine-tune the Motion Generator, further enhancing its capability. Experiments on both text-to-motion and music-to-dance generation tasks demonstrate that our framework achieves state-of-the-art motion generation quality while improving physical plausibility drastically., Comment: 15 pages, 6 figures

We show that a domain in $\mathbb{R}^3$ with the trace of a 3D Brownian motion removed almost surely satisfies the boundary Harnack principle (BHP). Then, we use it to prove that the intersection exponents for 3D Brownian motion are analytic., Comment: 49 pages, 5 figures

Humans can utilize techniques to quickly acquire knowledge from specific materials in advance, such as creating self-assessment questions, enabling us to achieving related tasks more efficiently. In contrast, large language models (LLMs) usually relies on retrieval-augmented generation to exploit knowledge materials in an instant manner, or requires external signals such as human preference data and stronger LLM annotations to conduct knowledge adaptation. To unleash the self-learning potential of LLMs, we propose KBAda, an approach designed for efficient adaptation to downstream tasks involving knowledge bases. Our method utilizes iterative training with self-annotated data such as Q&A pairs and revision suggestions, enabling the model to grasp the knowledge content efficiently. Experimental results on multiple datasets demonstrate the effectiveness of our approach, significantly boosting model performance in downstream tasks that require specific knowledge at a low cost. Notably, our approach achieves over 90% of the performance improvement that can be obtained by using GPT-4-turbo annotation, while relying entirely on self-supervision. We release our experimental data, models, and process analyses to the community for further exploration (https://github.com/thunlp/KBAda).

While high-quality texture maps are essential for realistic 3D asset rendering, few studies have explored learning directly in the texture space, especially on large-scale datasets. In this work, we depart from the conventional approach of relying on pre-trained 2D diffusion models for test-time optimization of 3D textures. Instead, we focus on the fundamental problem of learning in the UV texture space itself. For the first time, we train a large diffusion model capable of directly generating high-resolution texture maps in a feed-forward manner. To facilitate efficient learning in high-resolution UV spaces, we propose a scalable network architecture that interleaves convolutions on UV maps with attention layers on point clouds. Leveraging this architectural design, we train a 700 million parameter diffusion model that can generate UV texture maps guided by text prompts and single-view images. Once trained, our model naturally supports various extended applications, including text-guided texture inpainting, sparse-view texture completion, and text-driven texture synthesis. Project page is at http://cvmi-lab.github.io/TEXGen/., Comment: Accepted to SIGGRAPH Asia Journal Article (TOG 2024)

Electrochemical processes play a crucial role in energy storage and conversion systems, yet their computational modeling remains a significant challenge. Accurately incorporating the effects of electric potential has been a central focus in theoretical electrochemistry. Although constant-potential ab initio molecular dynamics (CP-AIMD) has provided valuable insights, it is limited by its substantial computational demands. Here, we introduce the Explicit Electric Potential Machine Learning Force Field (EEP-MLFF) model. Our model integrates the electric potential as an explicit input parameter along with the atom-centered descriptors in the atomic neural network. This approach enables the evaluation of nuclear forces under arbitrary electric potentials, thus facilitating molecular dynamics simulations at a specific potential. By applying the proposed machine learning method to the Cu/1T$^{\prime}$-MoS$_{2}$ system, molecular dynamics simulations reveal that the potential-modulated Cu atom migration and aggregation lead to the formation of small steric Cu clusters (Single Clusters, SCs) at potentials below -0.1 V. The morphological transformations of adsorbed Cu atoms are elucidated through electronic structure analyses, which demonstrates that both Cu-S and Cu-Cu bonding can be effectively tuned by the applied electric potential. Our findings present an opportunity for the convenient manufacture of single metal cluster catalysts through potential modulation. Moreover, this theoretical framework facilitates the exploration of potential-regulated processes and helps investigate the mechanisms of electrochemical reactions., Comment: 17 pages, 6 figures

We investigate phase transitions of the ground state (GS) of spin-1 Bose-Einstein condensates under the combined action of the spin-orbit coupling (SOC) and gradient magnetic field. Introducing appropariate raising and lowering operators, we exactly solve the linear system. Analyzing the obtained energy spectrum, we conclude that simultaneous variation of the magnetic-field gradient and SOC strength leads to the transition of excited states into the GS. As a result, any excited state can transition to the GS, at appropriate values of the system's parameters. The nonlinear system is solved numerically, showing that the GS phase transition, similar to the one in the linear system, still exists under the action of the repulsive nonlinearity. In the case of weak attraction, a mixed state appears near the GS transition point, while the GS transitions into an edge state under the action of strong attractive interaction., Comment: to be published in Physical Review A

The symbiosis of strong interactions, flat bands, topology and symmetry has led to the discovery of exotic phases of matter, including fractional Chern insulators, correlated moir\'e topological superconductors, and Dirac and Weyl semimetals. Correlated metals, such as those present in Kondo lattices, rely on the screening of local moments by a sea of non-magnetic conduction electrons. Here, we report on a unique topological Kondo lattice compound, CeCo2P2, where the Kondo effect - whose existence under the magnetic Co phase is protected by PT symmetry - coexists with antiferromagnetic order emerging from the flat bands associated with the Co atoms. Remarkably, this is the only known Kondo lattice compound where magnetic order occurs in non-heavy electrons, and puzzlingly, at a temperature significantly higher than that of the Kondo effect. Furthermore, at low temperatures, the emergence of the Kondo effect, in conjunction with a glide-mirror-z symmetry, results in a nodal line protected by bulk topology near the Fermi energy. These unusual properties, arising from the interplay between itinerant and correlated electrons from different constituent elements, lead to novel quantum phases beyond the celebrated topological Kondo insulators and Weyl Kondo semimetals. CeCo2P2 thus provides an ideal platform for investigating narrow bands, topology, magnetism, and the Kondo effect in strongly correlated electron systems., Comment: 17pages,4 figures

We present the application of the image coaddition algorithm, Up-sampling and PSF Deconvolution Coaddition (UPDC), for stacking multiple exposure images captured by the James Webb Space Telescope (JWST) Near-Infrared Camera (NIRCam). By addressing the point spread function (PSF) effect, UPDC provides visually enhanced and sharper images. Furthermore, the anti-aliasing and super-resolution capabilities of UPDC make it easier to deblend sources overlapped on images, yielding a higher accuracy of aperture photometry. We apply this algorithm to the SMACS J0723 imaging data. Comparative analysis with the Drizzle algorithm demonstrates significant improvements in detecting faint sources, achieving accurate photometry, and effectively deblending (super-resolution) closely packed sources. {As a result, we have newly detected a pair of close binary stars that were previously unresolvable in the original exposures or the Drizzled image.} These improvements significantly benefit various scientific projects conducted by JWST. The resulting dataset, named "UPdec-Webb", can be accessible through the official website of the Chinese Virtual Observatory (ChinaVO)., Comment: 21 pages, 18 figures, accepted for publication in ApJS

This work proposes a data-driven modeling and the corresponding hybrid motion control framework for unmanned and automated operation of industrial heavy-load hydraulic manipulator. Rather than the direct use of a neural network black box, we construct a reversible nonlinear model by using multilayer perceptron to approximate dynamics in the physical integrator chain system after reversible transformations. The reversible nonlinear model is trained offline using supervised learning techniques, and the data are obtained from simulations or experiments. Entire hybrid motion control framework consists of the model inversion controller that compensates for the nonlinear dynamics and proportional-derivative controller that enhances the robustness. The stability is proved with Lyapunov theory. Co-simulation and Experiments show the effectiveness of proposed modeling and hybrid control framework. With a commercial 39-ton class hydraulic excavator for motion control tasks, the root mean square error of trajectory tracking error decreases by at least 50\% compared to traditional control methods. In addition, by analyzing the system model, the proposed framework can be rapidly applied to different control plants.

Molecule discovery is a pivotal research field, impacting everything from the medicines we take to the materials we use. Recently, Large Language Models (LLMs) have been widely adopted in molecule understanding and generation, yet the alignments between molecules and their corresponding captions remain a significant challenge. Previous endeavours often treat the molecule as a general SMILES string or molecular graph, neglecting the fine-grained alignments between the molecular sub-structures and the descriptive textual phrases, which are crucial for accurate and explainable predictions. In this case, we introduce MolReFlect, a novel teacher-student framework designed to contextually perform the molecule-caption alignments in a fine-grained way. Our approach initially leverages a larger teacher LLM to label the detailed alignments by directly extracting critical phrases from molecule captions or SMILES strings and implying them to corresponding sub-structures or characteristics. To refine these alignments, we propose In-Context Selective Reflection, which retrieves previous extraction results as context examples for teacher LLM to reflect and lets a smaller student LLM select from in-context reflection and previous extraction results. Finally, we enhance the learning process of the student LLM through Chain-of-Thought In-Context Molecule Tuning, integrating the fine-grained alignments and the reasoning processes within the Chain-of-Thought format. Our experimental results demonstrate that MolReFlect enables LLMs like Mistral-7B to significantly outperform the previous baselines, achieving SOTA performance on the ChEBI-20 dataset. This advancement not only enhances the generative capabilities of LLMs in the molecule-caption translation task, but also contributes to a more explainable framework., Comment: 22 pages, 12 figures

Video Anomaly Detection (VAD) aims to automatically analyze spatiotemporal patterns in surveillance videos collected from open spaces to detect anomalous events that may cause harm without physical contact. However, vision-based surveillance systems such as closed-circuit television often capture personally identifiable information. The lack of transparency and interpretability in video transmission and usage raises public concerns about privacy and ethics, limiting the real-world application of VAD. Recently, researchers have focused on privacy concerns in VAD by conducting systematic studies from various perspectives including data, features, and systems, making Privacy-Preserving Video Anomaly Detection (P2VAD) a hotspot in the AI community. However, current research in P2VAD is fragmented, and prior reviews have mostly focused on methods using RGB sequences, overlooking privacy leakage and appearance bias considerations. To address this gap, this article systematically reviews the progress of P2VAD for the first time, defining its scope and providing an intuitive taxonomy. We outline the basic assumptions, learning frameworks, and optimization objectives of various approaches, analyzing their strengths, weaknesses, and potential correlations. Additionally, we provide open access to research resources such as benchmark datasets and available code. Finally, we discuss key challenges and future opportunities from the perspectives of AI development and P2VAD deployment, aiming to guide future work in the field., Comment: 19 pages, 6 figures

This paper introduces the Global Challenge for Safe and Secure Large Language Models (LLMs), a pioneering initiative organized by AI Singapore (AISG) and the CyberSG R&D Programme Office (CRPO) to foster the development of advanced defense mechanisms against automated jailbreaking attacks. With the increasing integration of LLMs in critical sectors such as healthcare, finance, and public administration, ensuring these models are resilient to adversarial attacks is vital for preventing misuse and upholding ethical standards. This competition focused on two distinct tracks designed to evaluate and enhance the robustness of LLM security frameworks. Track 1 tasked participants with developing automated methods to probe LLM vulnerabilities by eliciting undesirable responses, effectively testing the limits of existing safety protocols within LLMs. Participants were challenged to devise techniques that could bypass content safeguards across a diverse array of scenarios, from offensive language to misinformation and illegal activities. Through this process, Track 1 aimed to deepen the understanding of LLM vulnerabilities and provide insights for creating more resilient models.

Large Language Models (LLMs) demonstrate enhanced capabilities and reliability by reasoning more, evolving from Chain-of-Thought prompting to product-level solutions like OpenAI o1. Despite various efforts to improve LLM reasoning, high-quality long-chain reasoning data and optimized training pipelines still remain inadequately explored in vision-language tasks. In this paper, we present Insight-V, an early effort to 1) scalably produce long and robust reasoning data for complex multi-modal tasks, and 2) an effective training pipeline to enhance the reasoning capabilities of multi-modal large language models (MLLMs). Specifically, to create long and structured reasoning data without human labor, we design a two-step pipeline with a progressive strategy to generate sufficiently long and diverse reasoning paths and a multi-granularity assessment method to ensure data quality. We observe that directly supervising MLLMs with such long and complex reasoning data will not yield ideal reasoning ability. To tackle this problem, we design a multi-agent system consisting of a reasoning agent dedicated to performing long-chain reasoning and a summary agent trained to judge and summarize reasoning results. We further incorporate an iterative DPO algorithm to enhance the reasoning agent's generation stability and quality. Based on the popular LLaVA-NeXT model and our stronger base MLLM, we demonstrate significant performance gains across challenging multi-modal benchmarks requiring visual reasoning. Benefiting from our multi-agent system, Insight-V can also easily maintain or improve performance on perception-focused multi-modal tasks.

In this paper, we introduce DINO-X, which is a unified object-centric vision model developed by IDEA Research with the best open-world object detection performance to date. DINO-X employs the same Transformer-based encoder-decoder architecture as Grounding DINO 1.5 to pursue an object-level representation for open-world object understanding. To make long-tailed object detection easy, DINO-X extends its input options to support text prompt, visual prompt, and customized prompt. With such flexible prompt options, we develop a universal object prompt to support prompt-free open-world detection, making it possible to detect anything in an image without requiring users to provide any prompt. To enhance the model's core grounding capability, we have constructed a large-scale dataset with over 100 million high-quality grounding samples, referred to as Grounding-100M, for advancing the model's open-vocabulary detection performance. Pre-training on such a large-scale grounding dataset leads to a foundational object-level representation, which enables DINO-X to integrate multiple perception heads to simultaneously support multiple object perception and understanding tasks, including detection, segmentation, pose estimation, object captioning, object-based QA, etc. Experimental results demonstrate the superior performance of DINO-X. Specifically, the DINO-X Pro model achieves 56.0 AP, 59.8 AP, and 52.4 AP on the COCO, LVIS-minival, and LVIS-val zero-shot object detection benchmarks, respectively. Notably, it scores 63.3 AP and 56.5 AP on the rare classes of LVIS-minival and LVIS-val benchmarks, both improving the previous SOTA performance by 5.8 AP. Such a result underscores its significantly improved capacity for recognizing long-tailed objects., Comment: Technical Report

Our work aims to reconstruct hand-object interactions from a single-view image, which is a fundamental but ill-posed task. Unlike methods that reconstruct from videos, multi-view images, or predefined 3D templates, single-view reconstruction faces significant challenges due to inherent ambiguities and occlusions. These challenges are further amplified by the diverse nature of hand poses and the vast variety of object shapes and sizes. Our key insight is that current foundational models for segmentation, inpainting, and 3D reconstruction robustly generalize to in-the-wild images, which could provide strong visual and geometric priors for reconstructing hand-object interactions. Specifically, given a single image, we first design a novel pipeline to estimate the underlying hand pose and object shape using off-the-shelf large models. Furthermore, with the initial reconstruction, we employ a prior-guided optimization scheme, which optimizes hand pose to comply with 3D physical constraints and the 2D input image content. We perform experiments across several datasets and show that our method consistently outperforms baselines and faithfully reconstructs a diverse set of hand-object interactions. Here is the link of our project page: https://lym29.github.io/EasyHOI-page/, Comment: Project page: https://lym29.github.io/EasyHOI-page/

We propose a novel ferroelectric switchable altermagnetism effect, the reversal of ferroelectric polarization is coupled to the switching of altermagnetic spin splitting. We demonstrate the design principles for the ferroelectric altermagnets and the additional symmetry constraints necessary for switching the altermagnetic spin splitting through flipping the electric polarization based on the state-of-the-art spin-group symmetry techniques. 22 ferroelectric altermagnets are found by screening through the 2001 experimental reported magnetic structures in the MAGNDATA database and 2 of them are identified as ferroelectric switchable altermagnets. Using the hybrid improper ferroelectric material [C(NH2)3]Cr(HCOO)3 as an example, we show how the altermagnetic spin splitting is tightly coupled to the ferroelectric polarization, providing an ideal platform for designing electric-field-controllable multiferroic devices. Finally, we find that such manipulation of altermagnetism can be detected by monitoring the physical quantities that are related to the non-vanishing Berry curvature dipole, such as the linearly polarized photogalvanic spin current., Comment: 6 pages, 4 figures

Multiple federated learning (FL) methods are proposed for traffic flow forecasting (TFF) to avoid heavy-transmission and privacy-leaking concerns resulting from the disclosure of raw data in centralized methods. However, these FL methods adopt offline learning which may yield subpar performance, when concept drift occurs, i.e., distributions of historical and future data vary. Online learning can detect concept drift during model training, thus more applicable to TFF. Nevertheless, the existing federated online learning method for TFF fails to efficiently solve the concept drift problem and causes tremendous computing and communication overhead. Therefore, we propose a novel method named Resource-Efficient Federated Online Learning (REFOL) for TFF, which guarantees prediction performance in a communication-lightweight and computation-efficient way. Specifically, we design a data-driven client participation mechanism to detect the occurrence of concept drift and determine clients' participation necessity. Subsequently, we propose an adaptive online optimization strategy, which guarantees prediction performance and meanwhile avoids meaningless model updates. Then, a graph convolution-based model aggregation mechanism is designed, aiming to assess participants' contribution based on spatial correlation without importing extra communication and computing consumption on clients. Finally, we conduct extensive experiments on real-world datasets to demonstrate the superiority of REFOL in terms of prediction improvement and resource economization.

We report measurements of the absolute branching fractions $\mathcal{B}(B_s^0 \to D_s^{\pm} X)$, $\mathcal{B}(B_s^0 \to D^0/\bar{D}^0 X)$, and $\mathcal{B}(B_s^0 \to D^{\pm} X)$, where the latter is measured for the first time. The results are based on a 121.4\,fb$^{-1}$ data sample collected at the $\Upsilon(10860)$ resonance by the Belle detector at the KEKB asymmetric-energy $e^+ e^-$ collider. We reconstruct one $B_s^0$ meson in $e^+e^- \to \Upsilon(10860) \to B_s^{*} \bar{B}_s^{*}$ events and measure yields of $D_s^+$, $D^0$, and $D^+$ mesons in the rest of the event. We obtain $\mathcal{B}(B_s^0 \to D_s^{\pm} X) = (68.6 \pm 7.2 \pm 4.0)\%$, $\mathcal{B}(B_s^0 \to D^0/\bar{D}^0 X) = (21.5 \pm 6.1 \pm 1.8)\%$, and $\mathcal{B}(B_s^0 \to D^{\pm} X) = (12.6 \pm 4.6 \pm 1.3)\%$, where the first uncertainty is statistical and the second is systematic. Averaging with previous Belle measurements gives $\mathcal{B}(B_s^0 \to D_s^{\pm} X) = (63.4 \pm 4.5 \pm 2.2)\%$ and $\mathcal{B}(B_s^0 \to D^0/\bar{D}^0 X) = (23.9 \pm 4.1 \pm 1.8)\%$. For the $B_s^0$ production fraction at the $\Upsilon(10860)$, we find $f_s = (21.4^{+1.5}_{-1.7})\%$., Comment: 23 pages, 9 figures, submitted to JHEP

Display advertising provides significant value to advertisers, publishers, and users. Traditional display advertising systems utilize a multi-stage architecture consisting of retrieval, coarse ranking, and final ranking. However, conventional retrieval methods rely on ID-based learning to rank mechanisms and fail to adequately utilize the content information of ads, which hampers their ability to provide diverse recommendation lists. To address this limitation, we propose leveraging the extensive world knowledge of LLMs. However, three key challenges arise when attempting to maximize the effectiveness of LLMs: "How to capture user interests", "How to bridge the knowledge gap between LLMs and advertising system", and "How to efficiently deploy LLMs". To overcome these challenges, we introduce a novel LLM-based framework called LLM Empowered Display ADvertisement REcommender system (LEADRE). LEADRE consists of three core modules: (1) The Intent-Aware Prompt Engineering introduces multi-faceted knowledge and designs intent-aware pairs that fine-tune LLMs to generate ads tailored to users' personal interests. (2) The Advertising-Specific Knowledge Alignment incorporates auxiliary fine-tuning tasks and Direct Preference Optimization (DPO) to align LLMs with ad semantic and business value. (3) The Efficient System Deployment deploys LEADRE in an online environment by integrating both latency-tolerant and latency-sensitive service. Extensive offline experiments demonstrate the effectiveness of LEADRE and validate the contributions of individual modules. Online A/B test shows that LEADRE leads to a 1.57% and 1.17% GMV lift for serviced users on WeChat Channels and Moments separately. LEADRE has been deployed on both platforms, serving tens of billions of requests each day.

Recent advances in foundation models have established scaling laws that enable the development of larger models to achieve enhanced performance, motivating extensive research into large-scale recommendation models. However, simply increasing the model size in recommendation systems, even with large amounts of data, does not always result in the expected performance improvements. In this paper, we propose a novel framework, Collaborative Ensemble Training Network (CETNet), to leverage multiple distinct models, each with its own embedding table, to capture unique feature interaction patterns. Unlike naive model scaling, our approach emphasizes diversity and collaboration through collaborative learning, where models iteratively refine their predictions. To dynamically balance contributions from each model, we introduce a confidence-based fusion mechanism using general softmax, where model confidence is computed via negation entropy. This design ensures that more confident models have a greater influence on the final prediction while benefiting from the complementary strengths of other models. We validate our framework on three public datasets (AmazonElectronics, TaobaoAds, and KuaiVideo) as well as a large-scale industrial dataset from Meta, demonstrating its superior performance over individual models and state-of-the-art baselines. Additionally, we conduct further experiments on the Criteo and Avazu datasets to compare our method with the multi-embedding paradigm. Our results show that our framework achieves comparable or better performance with smaller embedding sizes, offering a scalable and efficient solution for CTR prediction tasks.

The ability to synthesize personalized group photos and specify the positions of each identity offers immense creative potential. While such imagery can be visually appealing, it presents significant challenges for existing technologies. A persistent issue is identity (ID) leakage, where injected facial features interfere with one another, resulting in low face resemblance, incorrect positioning, and visual artifacts. Existing methods suffer from limitations such as the reliance on segmentation models, increased runtime, or a high probability of ID leakage. To address these challenges, we propose ID-Patch, a novel method that provides robust association between identities and 2D positions. Our approach generates an ID patch and ID embeddings from the same facial features: the ID patch is positioned on the conditional image for precise spatial control, while the ID embeddings integrate with text embeddings to ensure high resemblance. Experimental results demonstrate that ID-Patch surpasses baseline methods across metrics, such as face ID resemblance, ID-position association accuracy, and generation efficiency. Project Page is: https://byteaigc.github.io/ID-Patch/, Comment: Project Page is: https://byteaigc.github.io/ID-Patch/

Evaluating the output of Large Language Models (LLMs) is one of the most critical aspects of building a performant compound AI system. Since the output from LLMs propagate to downstream steps, identifying LLM errors is crucial to system performance. A common task for LLMs in AI systems is tool use. While there are several benchmark environments for evaluating LLMs on this task, they typically only give a success rate without any explanation of the failure cases. To solve this problem, we introduce SpecTool, a new benchmark to identify error patterns in LLM output on tool-use tasks. Our benchmark data set comprises of queries from diverse environments that can be used to test for the presence of seven newly characterized error patterns. Using SPECTOOL , we show that even the most prominent LLMs exhibit these error patterns in their outputs. Researchers can use the analysis and insights from SPECTOOL to guide their error mitigation strategies.

Investigating physical models with photonic synthetic dimensions has been generating great interest in vast fields of science. The rapid developing thin-film lithium niobate (TFLN) platform, for its numerous advantages including high electro-optic coefficient and scalability, is well compatible with the realization of synthetic dimensions in the frequency together with spatial domain. While coupling resonators with fixed beam splitters is a common experimental approach, it often lacks tunability and limits coupling between adjacent lattices to sites occupying the same frequency domain positions. Here, on the contrary, we conceive the resonator arrays connected by electro-optic tunable Mach-Zehnder interferometers in our configuration instead of fixed beam splitters. By applying bias voltage and RF modulation on the interferometers, our design extends such coupling to long-range scenario and allows for continuous tuning on each coupling strength and synthetic effective magnetic flux. Therefore, our design enriches controllable coupling types that are essential for building programmable lattice networks and significantly increases versatility. As the example, we experimentally fabricate a two-resonator prototype on the TFLN platform, and on this single chip we realize well-known models including tight-binding lattices, topological Hall ladder and Creutz ladder. We directly observe the band structures in the quasi-momentum space and important phenomena such as spin-momentum locking and the Aharonov-Bohm cage effect. These results demonstrate the potential for convenient simulations of more complex models in our configuration.

The autobidding system generates huge revenue for advertising platforms, garnering substantial research attention. Existing studies in autobidding systems focus on designing Autobidding Incentive Compatible (AIC) mechanisms, where the mechanism is Incentive Compatible (IC) under ex ante expectations. However, upon deploying AIC mechanisms in advertising platforms, we observe a notable deviation between the actual auction outcomes and these expectations during runtime, particularly in the scene with few clicks (sparse-click). This discrepancy undermines truthful bidding among advertisers in AIC mechanisms, especially for risk-averse advertisers who are averse to outcomes that do not align with the expectations. To address this issue, we propose a mechanism, Decoupled First-Price Auction (DFP), that retains its IC property even during runtime. DFP dynamically adjusts the payment based on real-time user conversion outcomes, ensuring that advertisers' realized utilities closely approximate their expected utilities during runtime. To realize the payment mechanism of DFP, we propose a PPO-based RL algorithm, with a meticulously crafted reward function. This algorithm dynamically adjusts the payment to fit DFP mechanism. We conduct extensive experiments leveraging real-world data to validate our findings.

Existing click-through rate (CTR) prediction works have studied the role of feature interaction through a variety of techniques. Each interaction technique exhibits its own strength, and solely using one type could constrain the model's capability to capture the complex feature relationships, especially for industrial large-scale data with enormous users and items. Recent research shows that effective CTR models often combine an MLP network with a dedicated feature interaction network in a two-parallel structure. However, the interplay and cooperative dynamics between different streams or branches remain under-researched. In this work, we introduce a novel Multi-Branch Cooperation Network (MBCnet) which enables multiple branch networks to collaborate with each other for better complex feature interaction modeling. Specifically, MBCnet consists of three branches: the Expert-based Feature Grouping and Crossing (EFGC) branch that promotes the model's memorization ability of specific feature fields, the low rank Cross Net branch and Deep branch to enhance both explicit and implicit feature crossing for improved generalization. Among branches, a novel cooperation scheme is proposed based on two principles: branch co-teaching and moderate differentiation. Branch co-teaching encourages well-learned branches to support poorly-learned ones on specific training samples. Moderate differentiation advocates branches to maintain a reasonable level of difference in their feature representations. The cooperation strategy improves learning through mutual knowledge sharing via co-teaching and boosts the discovery of diverse feature interactions across branches. Extensive experiments on large-scale industrial datasets and online A/B test demonstrate MBCnet's superior performance, delivering a 0.09 point increase in CTR, 1.49% growth in deals, and 1.62% rise in GMV. Core codes will be released soon., Comment: 10 pages

The standing waves existed in radio telescope data are primarily due to reflections among the instruments, which significantly impact the spectrum quality of the Five-hundred-meter Aperture Spherical radio Telescope (FAST). Eliminating these standing waves for FAST is challenging given the constant changes in their phases and amplitudes. Over a ten-second period, the phases shift by 18$^{\circ}$ while the amplitudes fluctuate by 6 mK. Thus, we developed the fast Fourier transform (FFT) filter method to eliminate these standing waves for every individual spectrum. The FFT filter can decrease the root mean square (RMS) from 3.2 to 1.15 times the theoretical estimate. Compared to other methods such as sine fitting and running median, the FFT filter achieves a median RMS of approximately 1.2 times the theoretical expectation and the smallest scatter at 12%. Additionally, the FFT filter method avoids the flux loss issue encountered with some other methods. The FFT is also efficient in detecting harmonic radio frequency interference (RFI). In the FAST data, we identified three distinct types of harmonic RFI, each with amplitudes exceeding 100 mK and intrinsic frequency periods of 8.1, 0.5, and 0.37 MHz, respectively. The FFT filter, proven as the most effective method, is integrated into the HI data calibration and imaging pipeline for FAST (HiFAST, https://hifast.readthedocs.io)., Comment: 16 pages, 12 figures; accepted by RAA

Direct preference learning offers a promising and computation-efficient beyond supervised fine-tuning (SFT) for improving code generation in coding large language models (LMs). However, the scarcity of reliable preference data is a bottleneck for the performance of direct preference learning to improve the coding accuracy of code LMs. In this paper, we introduce \underline{\textbf{D}}irect Preference Learning with Only \underline{\textbf{S}}elf-Generated \underline{\textbf{T}}ests and \underline{\textbf{C}}ode (DSTC), a framework that leverages only self-generated code snippets and tests to construct reliable preference pairs such that direct preference learning can improve LM coding accuracy without external annotations. DSTC combines a minimax selection process and test-code concatenation to improve preference pair quality, reducing the influence of incorrect self-generated tests and enhancing model performance without the need for costly reward models. When applied with direct preference learning methods such as Direct Preference Optimization (DPO) and Kahneman-Tversky Optimization (KTO), DSTC yields stable improvements in coding accuracy (pass@1 score) across diverse coding benchmarks, including HumanEval, MBPP, and BigCodeBench, demonstrating both its effectiveness and scalability for models of various sizes. This approach autonomously enhances code generation accuracy across LLMs of varying sizes, reducing reliance on expensive annotated coding datasets.

Large language models (LLMs) have demonstrated impressive success in a wide range of natural language processing (NLP) tasks due to their extensive general knowledge of the world. Recent works discovered that the performance of LLMs is heavily dependent on the input prompt. However, prompt engineering is usually done manually in a trial-and-error fashion, which can be labor-intensive and challenging in order to find the optimal prompts. To address these problems and unleash the utmost potential of LLMs, we propose a novel LLMs-agnostic framework for prompt optimization, namely GRL-Prompt, which aims to automatically construct optimal prompts via reinforcement learning (RL) in an end-to-end manner. To provide structured action/state representation for optimizing prompts, we construct a knowledge graph (KG) that better encodes the correlation between the user query and candidate in-context examples. Furthermore, a policy network is formulated to generate the optimal action by selecting a set of in-context examples in a rewardable order to construct the prompt. Additionally, the embedding-based reward shaping is utilized to stabilize the RL training process. The experimental results show that GRL-Prompt outperforms recent state-of-the-art methods, achieving an average increase of 0.10 in ROUGE-1, 0.07 in ROUGE-2, 0.07 in ROUGE-L, and 0.05 in BLEU.

We have recently witnessed that ``Intelligence" and `` Compression" are the two sides of the same coin, where the language large model (LLM) with unprecedented intelligence is a general-purpose lossless compressor for various data modalities. This attribute particularly appeals to the lossless image compression community, given the increasing need to compress high-resolution images in the current streaming media era. Consequently, a spontaneous envision emerges: Can the compression performance of the LLM elevate lossless image compression to new heights? However, our findings indicate that the naive application of LLM-based lossless image compressors suffers from a considerable performance gap compared with existing state-of-the-art (SOTA) codecs on common benchmark datasets. In light of this, we are dedicated to fulfilling the unprecedented intelligence (compression) capacity of the LLM for lossless image compression tasks, thereby bridging the gap between theoretical and practical compression performance. Specifically, we propose P$^{2}$-LLM, a next-pixel prediction-based LLM, which integrates various elaborated insights and methodologies, \textit{e.g.,} pixel-level priors, the in-context ability of LLM, and a pixel-level semantic preservation strategy, to enhance the understanding capacity of pixel sequences for better next-pixel predictions. Extensive experiments on benchmark datasets demonstrate that P$^{2}$-LLM can beat SOTA classical and learned codecs.

We investigate the spin-density wave (SDW) behavior and the potential for superconductivity (SC) in La$_4$Ni$_3$O$_{10}$ under ambient pressure using a multi-orbital random-phase approximation (RPA). Starting with a twelve-orbital tight-binding model derived from density functional theory (DFT) calculations, we explore the influence of Hubbard interactions on SDW formation. Our analysis reveals a stripe-like SDW characterized by an incommensurate wave vector, $Q\approx(\pm 0.7\pi,0)$, suggesting a possible density wave instability in agreement with recent experiments. This configuration is driven by nesting of outer-layer Ni $d_{z^2}$ orbitals and exhibits interlayer antiferromagnetic ordering between the top and bottom NiO layers, with the middle layer serving as a node. We demonstrate that the Hund's coupling $J_H$ is the primary driver of the observed SDW. While superconductivity is absent in the undoped system under ambient pressure, it becomes attainable with appropriate hole doping ($\delta=-0.4$), resulting in a SC gap structure similar to the high-pressure phase. Our study identifies the specific conditions for realizing the ambient pressure stripe density wave: $J_H>0.16U$. Additionally, when doping leads to sufficient nesting at (0,0), the system's magnetic fluctuations transition into a stable Neel-type antiferromagnetic state, analogous to the high-pressure case., Comment: 6 pages, 7 figures

Deploying robots in open-world environments involves complex tasks characterized by long sequences and rich interactions, necessitating efficient transfer of robotic skills across diverse and complex scenarios. To address this challenge, we propose a skill library framework based on knowledge graphs, which endows robots with high-level skill awareness and spatial semantic understanding. The framework hierarchically organizes operational knowledge by constructing a "task graph" and a "scene graph" to represent task and scene semantic information, respectively. We introduce a "state graph" to facilitate interaction between high-level task planning and low-level scene information. Furthermore, we propose a hierarchical transfer framework for operational skills. At the task level, the framework integrates contextual learning and chain-of-thought prompting within a four-stage prompt paradigm, leveraging large language models' (LLMs) reasoning and generalization capabilities to achieve task-level subtask sequence transfer. At the motion level, an adaptive trajectory transfer method is developed using the A* algorithm and the skill library, enabling motion-level adaptive trajectory transfer. At the physical level, we introduce an adaptive contour extraction and posture perception method based on tactile perception. This method dynamically obtains high-precision contour and posture information from visual-tactile texture data and adjusts transferred skills, such as contact positions and postures, to ensure effectiveness in new environments. Experimental results validate the effectiveness of the proposed methods. Project website:https://github.com/MingchaoQi/skill_transfer

Using $e^+e^-$ collision data corresponding to an integrated luminosity of 19 fb$^{-1}$ collected by the BESIII detector at center-of-mass energies ranging from 4.13 to 4.70 GeV, we report the first evidence for a new excited $\Omega^{-}$ hyperon, the $\Omega^*(2109)^{-}$, through the process $e^+ e^- \to \Omega^*(2109)^{-} \bar{\Omega}^{+} +c.c.$ with a significance of 3.7 $\sigma$. The mass and width of $\Omega^*(2109)^{-}$ are measured to be $2108.8 \pm 5.5_{\rm stat} \pm 1.5_{\rm syst} {\rm MeV}/c^{2}$ and $21.6 \pm 17.7_{\rm stat} \pm 9.4_{\rm syst} {\rm MeV}$, respectively. We also present evidence for production of the $\Omega^*(2012)^{-}$ in the process $e^+ e^- \to \Omega^*(2012)^{-} \bar{\Omega}^{+} +c.c.$ with a significance of 3.7 $\sigma$., Comment: 8 pages, 2 figures

In recent years, the rapid growth of security vulnerabilities poses great challenges to tracing and managing them. For example, it was reported that the NVD database experienced significant delays due to the shortage of maintainers. Such delay creates challenges for third-party security personnel (e.g., administrators) to trace the information related to the CVE. To help security personnel trace a vulnerability patch, we build a retrieval system that automatically retrieves the patch in the repository. Inspired by existing work on explainable machine learning, we ask the following research question: can explanations help security maintainers make decisions in patch tracing? First, we investigate using LIME (a widely used explainable machine learning method) to highlight the rationale tokens in the commit message and code. In addition, we propose an explanation method called TfIdf-Highlight, which leverages the Tf-Idf statistics to select the most informative words in the repository and the dataset. We evaluate the effectiveness of highlighting using two experiments. First, we compare LIME and TfIdf-Highlight using a faithfulness score (i.e., sufficiency and comprehensiveness) defined for ranking. We find that TfIdf-Highlight significantly outperforms LIME's sufficiency scores by 15\% and slightly outperforms the comprehensiveness scores. Second, we conduct a blind human labeling experiment by asking the annotators to guess the patch under 3 settings (TfIdf-Highlight, LIME, and no highlight). We find that the helpfulness score for TfIdf-Highlight is higher than LIME while the labeling accuracies of LIME and TfIdf-Highlight are similar. Nevertheless, highlighting does not improve the accuracy over non-highlighting.

Shape assembly is a ubiquitous task in daily life, integral for constructing complex 3D structures like IKEA furniture. While significant progress has been made in developing autonomous agents for shape assembly, existing datasets have not yet tackled the 4D grounding of assembly instructions in videos, essential for a holistic understanding of assembly in 3D space over time. We introduce IKEA Video Manuals, a dataset that features 3D models of furniture parts, instructional manuals, assembly videos from the Internet, and most importantly, annotations of dense spatio-temporal alignments between these data modalities. To demonstrate the utility of IKEA Video Manuals, we present five applications essential for shape assembly: assembly plan generation, part-conditioned segmentation, part-conditioned pose estimation, video object segmentation, and furniture assembly based on instructional video manuals. For each application, we provide evaluation metrics and baseline methods. Through experiments on our annotated data, we highlight many challenges in grounding assembly instructions in videos to improve shape assembly, including handling occlusions, varying viewpoints, and extended assembly sequences., Comment: NeurIPS 2024 Datasets and Benchmarks Track

Differentiable rendering techniques have recently shown promising results for free-viewpoint video synthesis of characters. However, such methods, either Gaussian Splatting or neural implicit rendering, typically necessitate per-subject optimization which does not meet the requirement of real-time rendering in an interactive application. We propose a generalizable Gaussian Splatting approach for high-resolution image rendering under a sparse-view camera setting. To this end, we introduce Gaussian parameter maps defined on the source views and directly regress Gaussian properties for instant novel view synthesis without any fine-tuning or optimization. We train our Gaussian parameter regression module on human-only data or human-scene data, jointly with a depth estimation module to lift 2D parameter maps to 3D space. The proposed framework is fully differentiable with both depth and rendering supervision or with only rendering supervision. We further introduce a regularization term and an epipolar attention mechanism to preserve geometry consistency between two source views, especially when neglecting depth supervision. Experiments on several datasets demonstrate that our method outperforms state-of-the-art methods while achieving an exceeding rendering speed., Comment: Journal extension of CVPR 2024,Project page:https://yaourtb.github.io/GPS-Gaussian+

Rhombohedral multilayer graphene has emerged as an extraordinary platform for investigating exotic quantum states, such as superconductivity and fractional quantum anomalous Hall effects, mainly due to the existence of topological surface flatbands. Despite extensive research efforts, a systematic spectroscopic investigation on the evolution of its electronic structure from thin layers to bulk remains elusive. Using state-of-the-art angle-resolved photoemission spectroscopy with submicron spatial resolution, we directly probe and trace the thickness evolution of the topological electronic structures of rhombohedral multilayer graphene. As the layer number increases, the gapped subbands transform into the 3D Dirac nodes that spirals in the momentum space; while the flatbands are constantly observed around Fermi level, and eventually evolve into the topological drumhead surface states. This unique thickness-dependent topological phase transition can be well captured by the 3D generalization of 1D Su-Schrieffer-Heeger chain in thin layers, to the topological Dirac nodal spiral semimetal in the bulk limit. Our findings establish a solid foundation for exploring the exotic quantum phases with nontrivial topology and correlation effects in rhombohedral multilayer graphene., Comment: 15 pages, 4 figures, under review

Oracle character recognition-an analysis of ancient Chinese inscriptions found on oracle bones-has become a pivotal field intersecting archaeology, paleography, and historical cultural studies. Traditional methods of oracle character recognition have relied heavily on manual interpretation by experts, which is not only labor-intensive but also limits broader accessibility to the general public. With recent breakthroughs in pattern recognition and deep learning, there is a growing movement towards the automation of oracle character recognition (OrCR), showing considerable promise in tackling the challenges inherent to these ancient scripts. However, a comprehensive understanding of OrCR still remains elusive. Therefore, this paper presents a systematic and structured survey of the current landscape of OrCR research. We commence by identifying and analyzing the key challenges of OrCR. Then, we provide an overview of the primary benchmark datasets and digital resources available for OrCR. A review of contemporary research methodologies follows, in which their respective efficacies, limitations, and applicability to the complex nature of oracle characters are critically highlighted and examined. Additionally, our review extends to ancillary tasks associated with OrCR across diverse disciplines, providing a broad-spectrum analysis of its applications. We conclude with a forward-looking perspective, proposing potential avenues for future investigations that could yield significant advancements in the field.

A new caesium (Cs) fountain clock NIM6 has been developed at the National Institute of Metrology (NIM) in China, for which a comprehensive uncertainty evaluation is presented. A three-dimensional magneto-optical trap (3D MOT) loading optical molasses is employed to obtain more cold atoms rapidly and efficiently with a tunable, uniform density distribution. A heat pipe surrounding the flight tube maintains a consistent and stable temperature within the interrogation region. Additionally, a Ramsey cavity with four azimuthally distribution feeds is utilized to mitigate distributed cavity phase shifts. The Cs fountain clock NIM6 achieves a short-term stability of 1.0x10-13 {\tau}-1/2 at high atomic density, and a typical overall fractional type-B uncertainty is estimated to be 2.3x10-16. Comparisons of frequency between the Cs fountain NIM6 and other Cs fountain Primary Frequency Standards (PFSs) through Coordinated Universal Time (UTC) have demonstrated an agreement within the stated uncertainties., Comment: 14 pages, 16 figures

The $C\!P$ asymmetry and branching fraction of the CKM-suppressed decay $B^+\!\to J\mskip -3mu/\mskip -2mu\psi\,\pi^+$ are precisely measured relative to the favoured decay $B^+\!\to J\mskip -3mu/\mskip -2mu\psi\,K^+$, using a sample of proton-proton collision data corresponding to an integrated luminosity of $5.4~\mathrm{fb}^{-1}$ recorded at center-of-mass energy of $13~\mathrm{TeV}$ during 2016--2018. The results of the $C\!P$ asymmetry difference and branching fraction ratio are \begin{align*} \Delta\mathcal{A}^{C\!P} &\equiv \mathcal{A}^{C\!P}(B^+ \to J\mskip -3mu/\mskip -2mu\psi\,\pi^+) - \mathcal{A}^{C\!P}(B^+ \to J\mskip -3mu/\mskip -2mu\psi\,K^+) = (1.29 \pm 0.49 \pm 0.08) \times 10^{-2}, \end{align*} \begin{equation*} \mathcal{R}_{\pi/K} \equiv \frac{\mathcal{B}(B^+ \!\to J\mskip -3mu/\mskip -2mu\psi\,\pi^+)}{\mathcal{B}(B^+ \!\to J\mskip -3mu/\mskip -2mu\psi\,K^+)} = (3.852 \pm 0.022 \pm 0.018) \times 10^{-2}. \end{equation*} where the first uncertainties are statistical and the second systematic. A combination with previous LHCb results based on data collected at $7$ and $8~\mathrm{TeV}$ in 2011 and 2012 yields $\Delta\mathcal{A}^{C\!P} = (1.42 \pm 0.43 \pm 0.08) \times 10^{-2}$ and $\mathcal{R}_{\pi/K} = (3.846 \pm 0.018 \pm 0.018) \times 10^{-2}$. The combined $\Delta\mathcal{A}^{C\!P}$ value deviates from zero by 3.2 standard deviations, providing the first evidence for direct $C\!P$ violation in the amplitudes of beauty decays to charmonium final states., Comment: 18 pages, 2 figures, no conference or journal information All figures and tables, along with machine-readable versions and any supplementary material and additional information, are available at https://lbfence.cern.ch/alcm/public/analysis/full-details/1623/ (LHCb public pages)

Decision Transformer (DT) can learn effective policy from offline datasets by converting the offline reinforcement learning (RL) into a supervised sequence modeling task, where the trajectory elements are generated auto-regressively conditioned on the return-to-go (RTG).However, the sequence modeling learning approach tends to learn policies that converge on the sub-optimal trajectories within the dataset, for lack of bridging data to move to better trajectories, even if the condition is set to the highest RTG.To address this issue, we introduce Diffusion-Based Trajectory Branch Generation (BG), which expands the trajectories of the dataset with branches generated by a diffusion model.The trajectory branch is generated based on the segment of the trajectory within the dataset, and leads to trajectories with higher returns.We concatenate the generated branch with the trajectory segment as an expansion of the trajectory.After expanding, DT has more opportunities to learn policies to move to better trajectories, preventing it from converging to the sub-optimal trajectories.Empirically, after processing with BG, DT outperforms state-of-the-art sequence modeling methods on D4RL benchmark, demonstrating the effectiveness of adding branches to the dataset without further modifications.

Large Language Models (LLMs) have demonstrated remarkable success across a wide range of tasks and domains. However, their performance in low-resource language translation, particularly when translating into these languages, remains underexplored. This gap poses significant challenges, as linguistic barriers hinder the cultural preservation and development of minority communities. To address this issue, this paper introduces a novel retrieval-based method that enhances translation quality for low-resource languages by focusing on key terms, which involves translating keywords and retrieving corresponding examples from existing data. To evaluate the effectiveness of this method, we conducted experiments translating from English into three low-resource languages: Cherokee, a critically endangered indigenous language of North America; Tibetan, a historically and culturally significant language in Asia; and Manchu, a language with few remaining speakers. Our comparison with the zero-shot performance of GPT-4o and LLaMA 3.1 405B, highlights the significant challenges these models face when translating into low-resource languages. In contrast, our retrieval-based method shows promise in improving both word-level accuracy and overall semantic understanding by leveraging existing resources more effectively.

Unsupervised sentence representation learning remains a critical challenge in modern natural language processing (NLP) research. Recently, contrastive learning techniques have achieved significant success in addressing this issue by effectively capturing textual semantics. Many such approaches prioritize the optimization using negative samples. In fields such as computer vision, hard negative samples (samples that are close to the decision boundary and thus more difficult to distinguish) have been shown to enhance representation learning. However, adapting hard negatives to contrastive sentence learning is complex due to the intricate syntactic and semantic details of text. To address this problem, we propose HNCSE, a novel contrastive learning framework that extends the leading SimCSE approach. The hallmark of HNCSE is its innovative use of hard negative samples to enhance the learning of both positive and negative samples, thereby achieving a deeper semantic understanding. Empirical tests on semantic textual similarity and transfer task datasets validate the superiority of HNCSE.

We present the results of our system for the CoMeDi Shared Task, which predicts majority votes (Subtask 1) and annotator disagreements (Subtask 2). Our approach combines model ensemble strategies with MLP-based and threshold-based methods trained on pretrained language models. Treating individual models as virtual annotators, we simulate the annotation process by designing aggregation measures that incorporate continuous similarity scores and discrete classification labels to capture both majority and disagreement. Additionally, we employ anisotropy removal techniques to enhance performance. Experimental results demonstrate the effectiveness of our methods, particularly for Subtask 2. Notably, we find that continuous similarity scores, even within the same model, align better with human disagreement patterns compared to aggregated discrete labels., Comment: 8 pages, 3 figures

Rapid development of artificial intelligence has drastically accelerated the development of scientific discovery. Trained with large-scale observation data, deep neural networks extract the underlying patterns in an end-to-end manner and assist human researchers with highly-precised predictions in unseen scenarios. The recent rise of Large Language Models (LLMs) and the empowered autonomous agents enable scientists to gain help through interaction in different stages of their research, including but not limited to literature review, research ideation, idea implementation, and academic writing. However, AI researchers instantiated by foundation model empowered agents with full-process autonomy are still in their infancy. In this paper, we study $\textbf{AI-Generated Science}$ (AIGS), where agents independently and autonomously complete the entire research process and discover scientific laws. By revisiting the definition of scientific research, we argue that $\textit{falsification}$ is the essence of both human research process and the design of an AIGS system. Through the lens of falsification, prior systems attempting towards AI-Generated Science either lack the part in their design, or rely heavily on existing verification engines that narrow the use in specialized domains. In this work, we propose Baby-AIGS as a baby-step demonstration of a full-process AIGS system, which is a multi-agent system with agents in roles representing key research process. By introducing FalsificationAgent, which identify and then verify possible scientific discoveries, we empower the system with explicit falsification. Experiments on three tasks preliminarily show that Baby-AIGS could produce meaningful scientific discoveries, though not on par with experienced human researchers. Finally, we discuss on the limitations of current Baby-AIGS, actionable insights, and related ethical issues in detail., Comment: Pre-print. 35 pages. Official website: https://agent-force.github.io/AIGS/