Search

Manifestly and logically displaying the line of reasoning from evidence to answer is significant to explainable question answering (QA). The entailment tree exhibits the lines structurally, which is different from the self-explanation principle in large-scale language models. Existing methods rarely consider the semantic association of sentences between and within hierarchies within the tree structure, which is prone to apparent mistakes in combinations. In this work, we propose an architecture of integrating the Hierarchical Semantics of sentences under the framework of Controller-Generator (HiSCG) to explain answers. The HiSCG designs a hierarchical mapping between hypotheses and facts, discriminates the facts involved in tree constructions, and optimizes single-step entailments. To the best of our knowledge, We are the first to notice hierarchical semantics of sentences between the same layer and adjacent layers to yield improvements. The proposed method achieves comparable performance on all three settings of the EntailmentBank dataset. The generalization results on two out-of-domain datasets also demonstrate the effectiveness of our method.

Large language model (LLM)-based AI delegates are increasingly utilized to act on behalf of users, assisting them with a wide range of tasks through conversational interfaces. Despite their advantages, concerns arise regarding the potential risk of privacy leaks, particularly in scenarios involving social interactions. While existing research has focused on protecting privacy by limiting the access of AI delegates to sensitive user information, many social scenarios require disclosing private details to achieve desired outcomes, necessitating a balance between privacy protection and disclosure. To address this challenge, we conduct a pilot study to investigate user preferences for AI delegates across various social relations and task scenarios, and then propose a novel AI delegate system that enables privacy-conscious self-disclosure. Our user study demonstrates that the proposed AI delegate strategically protects privacy, pioneering its use in diverse and dynamic social interactions.

Existing 3D mask learning methods encounter performance bottlenecks under limited data, and our objective is to overcome this limitation. In this paper, we introduce a triple point masking scheme, named TPM, which serves as a scalable framework for pre-training of masked autoencoders to achieve multi-mask learning for 3D point clouds. Specifically, we augment the baselines with two additional mask choices (i.e., medium mask and low mask) as our core insight is that the recovery process of an object can manifest in diverse ways. Previous high-masking schemes focus on capturing the global representation but lack the fine-grained recovery capability, so that the generated pre-trained weights tend to play a limited role in the fine-tuning process. With the support of the proposed TPM, available methods can exhibit more flexible and accurate completion capabilities, enabling the potential autoencoder in the pre-training stage to consider multiple representations of a single 3D object. In addition, an SVM-guided weight selection module is proposed to fill the encoder parameters for downstream networks with the optimal weight during the fine-tuning stage, maximizing linear accuracy and facilitating the acquisition of intricate representations for new objects. Extensive experiments show that the four baselines equipped with the proposed TPM achieve comprehensive performance improvements on various downstream tasks. Our code and models are available at https://github.com/liujia99/TPM.

As a point defect with unique spin and optical properties, nitrogen-vacancy (NV) center in diamond has attracted much attention in the fields of quantum sensing, quantum simulation, and quantum networks. The optical properties of an NV center are crucial for all these quantum applications. However, NV centers fabricated by destructive methods such as electron irradiation or ion implantation usually exhibit poor optical coherence. In this work, we demonstrate a non-destructive method to fabricate optically coherent NV centers. High-purity single crystal diamonds are annealed under high pressure and high temperature (1700 $^{\circ}$C, 5.5 GPa), and individually resolvable NV centers with narrow PLE linewidth (<100 MHz) are produced. The high-pressure condition prevents the conversion of diamond to graphite during high-temperature annealing, significantly expanding the parameter space for creating high-performance artificial defects for quantum information science. These findings deepen our understanding of NV center formation in diamond and have implications for the optimization of color centers in solids, including silicon carbide and hexagonal boron nitride., Comment: 11 pages,4 figures

One of the major challenges for the agricultural industry today is the uncertainty in manual labor availability and the associated cost. Automated flower and fruit density estimation, localization, and counting could help streamline harvesting, yield estimation, and crop-load management strategies such as flower and fruitlet thinning. This article proposes a deep regression-based network, AgRegNet, to estimate density, count, and location of flower and fruit in tree fruit canopies without explicit object detection or polygon annotation. Inspired by popular U-Net architecture, AgRegNet is a U-shaped network with an encoder-to-decoder skip connection and modified ConvNeXt-T as an encoder feature extractor. AgRegNet can be trained based on information from point annotation and leverages segmentation information and attention modules (spatial and channel) to highlight relevant flower and fruit features while suppressing non-relevant background features. Experimental evaluation in apple flower and fruit canopy images under an unstructured orchard environment showed that AgRegNet achieved promising accuracy as measured by Structural Similarity Index (SSIM), percentage Mean Absolute Error (pMAE) and mean Average Precision (mAP) to estimate flower and fruit density, count, and centroid location, respectively. Specifically, the SSIM, pMAE, and mAP values for flower images were 0.938, 13.7%, and 0.81, respectively. For fruit images, the corresponding values were 0.910, 5.6%, and 0.93. Since the proposed approach relies on information from point annotation, it is suitable for sparsely and densely located objects. This simplified technique will be highly applicable for growers to accurately estimate yields and decide on optimal chemical and mechanical flower thinning practices.

A search for the very rare $B^{*0}\to\mu^+\mu^-$ and $B_{s}^{*0}\to\mu^+\mu^-$ decays is conducted by analysing the $B_c^+\to \pi^+\mu^+\mu^-$ process. The analysis uses proton-proton collision data collected with the LHCb detector between 2011 and 2018, corresponding to an integrated luminosity of 9$\text{\,fb}^{-1}$. The signal signatures correspond to simultaneous peaks in the $\mu^+\mu^-$ and $\pi^+\mu^+\mu^-$ invariant masses. No evidence for an excess of events over background is observed for either signal decay mode. Upper limits at the $90\%$ confidence level are set on the branching fractions relative to that for $B_c^+\to J\mskip -3mu/\mskip -2mu\psi\pi^+$ decays, \begin{align*} {\cal R}_{B^{*0}(\mu^+\mu^-)\pi^+/J\mskip -3mu/\mskip -2mu\psi\pi^+} &< 3.8\times 10^{-5}\ \text{ and } {\cal R}_{B_{s}^{*0}(\mu^+\mu^-)\pi^+/J\mskip -3mu/\mskip -2mu\psi\pi^+} &< 5.0\times 10^{-5}\,. \end{align*}, Comment: 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/1796/ (LHCb public pages)

Fast Radio Bursts (FRBs) are millisecond-duration events detected from beyond the Milky Way. FRB emission characteristics favor highly magnetized neutron stars, or magnetars, as the sources, as evidenced by FRB-like bursts from a galactic magnetar, and the star-forming nature of FRB host galaxies. However, the processes that produce FRB sources remain unknown. Although galactic magnetars are often linked to core-collapse supernovae (CCSNe), it's uncertain what determines which supernovae result in magnetars. The galactic environments of FRB sources can be harnessed to probe their progenitors. Here, we present the stellar population properties of 30 FRB host galaxies discovered by the Deep Synoptic Array. Our analysis shows a significant deficit of low-mass FRB hosts compared to the occurrence of star-formation in the universe, implying that FRBs are a biased tracer of star-formation, preferentially selecting massive star-forming galaxies. This bias may be driven by galaxy metallicity, which is positively correlated with stellar mass. Metal-rich environments may favor the formation of magnetar progenitors through stellar mergers, as higher metallicity stars are less compact and more likely to fill their Roche lobes, leading to unstable mass transfer. Although massive stars do not have convective interiors to generate strong magnetic fields by dynamo, merger remnants are thought to have the requisite internal magnetic-field strengths to result in magnetars. The preferential occurrence of FRBs in massive star-forming galaxies suggests that CCSN of merger remnants preferentially forms magnetars., Comment: Accepted for publication in Nature. The final version will be published by the journal

Multi-view clustering aims to study the complementary information across views and discover the underlying structure. For solving the relatively high computational cost for the existing approaches, works based on anchor have been presented recently. Even with acceptable clustering performance, these methods tend to map the original representation from multiple views into a fixed shared graph based on the original dataset. However, most studies ignore the discriminative property of the learned anchors, which ruin the representation capability of the built model. Moreover, the complementary information among anchors across views is neglected to be ensured by simply learning the shared anchor graph without considering the quality of view-specific anchors. In this paper, we propose discriminative anchor learning for multi-view clustering (DALMC) for handling the above issues. We learn discriminative view-specific feature representations according to the original dataset and build anchors from different views based on these representations, which increase the quality of the shared anchor graph. The discriminative feature learning and consensus anchor graph construction are integrated into a unified framework to improve each other for realizing the refinement. The optimal anchors from multiple views and the consensus anchor graph are learned with the orthogonal constraints. We give an iterative algorithm to deal with the formulated problem. Extensive experiments on different datasets show the effectiveness and efficiency of our method compared with other methods., Comment: This work has been accepted by TMM

Classical and quantum states can be distinguished by entanglement entropy, which can be viewed as a measure of quantum resources. Entanglement entropy also plays a pivotal role in understanding computational complexity in simulating quantum systems. However, stabilizer states formed solely by Clifford gates can be efficiently simulated with the tableau algorithm according to the Gottesman-Knill theorem, although they can host large entanglement entropy. In this work, we introduce the concept of non-stabilizerness entanglement entropy which is basically the minimum residual entanglement entropy for a quantum state by excluding the contribution from Clifford circuits. It can serve as a new practical and better measure of difficulty in the classical simulation of quantum many-body systems. We discuss why it is a better criterion than previously proposed metrics such as Stabilizer R\'enyi Entropy. We also show numerical results of non-stabilizerness entanglement entropy with concrete quantum many-body models. The concept of non-stabilizerness entanglement entropy expands our understanding of the ``hardness`` in the classical simulation of quantum many-body systems.

Large language models (LLMs) have achieved remarkable advancements in natural language processing, showcasing exceptional performance across various tasks. However, the expensive memory and computational requirements present significant challenges for their practical deployment. Low-bit quantization has emerged as a critical approach to mitigate these challenges by reducing the bit-width of model parameters, activations, and gradients, thus decreasing memory usage and computational demands. This paper presents a comprehensive survey of low-bit quantization methods tailored for LLMs, covering the fundamental principles, system implementations, and algorithmic strategies. An overview of basic concepts and new data formats specific to low-bit LLMs is first introduced, followed by a review of frameworks and systems that facilitate low-bit LLMs across various hardware platforms. Then, we categorize and analyze techniques and toolkits for efficient low-bit training and inference of LLMs. Finally, we conclude with a discussion of future trends and potential advancements of low-bit LLMs. Our systematic overview from basic, system, and algorithm perspectives can offer valuable insights and guidelines for future works to enhance the efficiency and applicability of LLMs through low-bit quantization., Comment: Ruihao Gong leads the overall organization of the survey, with Yifu Ding and Jinyang Du contributing to Sections 2 and 3. Xingyu Zheng is responsible for authoring Section 4, while Chengtao Lv and Zining Wang collaborate on Section 5. Haotong Qin, Jinyang Guo, Michele Magno, and Xianglong Liu provide guidance during the whole process and assist in refining the final manuscript

Integrating aerial imagery-based scene generation into applications like autonomous driving and gaming enhances realism in 3D environments, but challenges remain in creating detailed content for occluded areas and ensuring real-time, consistent rendering. In this paper, we introduce Skyeyes, a novel framework that can generate photorealistic sequences of ground view images using only aerial view inputs, thereby creating a ground roaming experience. More specifically, we combine a 3D representation with a view consistent generation model, which ensures coherence between generated images. This method allows for the creation of geometrically consistent ground view images, even with large view gaps. The images maintain improved spatial-temporal coherence and realism, enhancing scene comprehension and visualization from aerial perspectives. To the best of our knowledge, there are no publicly available datasets that contain pairwise geo-aligned aerial and ground view imagery. Therefore, we build a large, synthetic, and geo-aligned dataset using Unreal Engine. Both qualitative and quantitative analyses on this synthetic dataset display superior results compared to other leading synthesis approaches. See the project page for more results: https://chaoren2357.github.io/website-skyeyes/.

The increasing number of dispersed EEG dataset publications and the advancement of large-scale Electroencephalogram (EEG) models have increased the demand for practical tools to manage diverse EEG datasets. However, the inherent complexity of EEG data, characterized by variability in content data, metadata, and data formats, poses challenges for integrating multiple datasets and conducting large-scale EEG model research. To tackle the challenges, this paper introduces EEGUnity, an open-source tool that incorporates modules of 'EEG Parser', 'Correction', 'Batch Processing', and 'Large Language Model Boost'. Leveraging the functionality of such modules, EEGUnity facilitates the efficient management of multiple EEG datasets, such as intelligent data structure inference, data cleaning, and data unification. In addition, the capabilities of EEGUnity ensure high data quality and consistency, providing a reliable foundation for large-scale EEG data research. EEGUnity is evaluated across 25 EEG datasets from different sources, offering several typical batch processing workflows. The results demonstrate the high performance and flexibility of EEGUnity in parsing and data processing. The project code is publicly available at github.com/Baizhige/EEGUnity.

The myelinated axons are widely present in both central and peripheral nervous systems. Its unique compact spiraling structure poses significant challenges to understanding its biological functions and developmental mechanisms. Conventionally, myelin is considered as an insulating layer to achieve saltatory conduction for the enhancement of the neural signal speed, which serves as the foundation of neuroscience. However, this insulating hypothesis is inadequate to account for various experimental observations, especially the long unmyelinated tract observed in the cortex. We here show non-random distributions in three ultrastructural features of myelin: the non-random spiraling directions, the localization preferences of myelin outer tongues, and the radial components along boundaries between oppositely spiraled myelin sheaths. These phenomena are predicted by a novel concept of myelin biological function, which we propose as the signal synchronization function. Our findings demonstrate that cytoplasmic channels within myelin may act as coiled inductors, facilitating electromagnetic induction between adjacent myelin sheaths, and thereby promoting signal synchronization between axons. This, in turn, explains the non-random ultrastructural features observed. We believe these insights lay the foundation for a new understanding of myelin inductive function.

Motivated by the recent discovery of superconductivity in twisted bilayer WSe$_2$, we theoretically investigate the pairing instability induced by Coulomb interaction via the Kohn-Luttinger mechanism. Within random phase approximation, we show that the Coulomb repulsion is strongly screened around the mori\'e length scale, resulting in attractive regimes that facilitate electron pairing in unconventional channels. By solving the linearized gap equation, we find that the maxima of the superconducting critical temperature occur at the van Hove singularities in the density of states, with magnitude on the order of $\sim$0.2 K, comparable with experimental observations. Moreover, we identify a phase transition from an $f$-wave-like pairing state to a doubly degenerate $d$-wave-like pairing state upon increasing the electric displacement field. These findings provide a purely electronic explanation of the superconducting state observed in twisted bilayer WSe$_2$, offering insights into the interplay between superconductivity and other correlated states in twisted transition metal dichalcogenides., Comment: 7+7 pages, 5+6 figures

In this paper, we investigate the generation of baryon asymmetry of the universe (BAU) following the first-order electroweak phase transition. Our study indicates that the presence of CP-violating operators can lead to the generation of the helical magnetic field, which further induces the growth of the lepton asymmetry and the generation of the BAU as the Universe cools down. This process can yield the observed BAU when the new physics scale is lower than 700 GeV., Comment: 24 pages, 11 figures, comments welcome!

Numerically the reconstructability of unknown parameters in inverse problems heavily relies on the chosen data. Therefore, it is crucial to design an experiment that yields data that is sensitive to the parameters. We approach this problem from the perspective of a least squares optimization, and examine the positivity of the Gauss-Newton Hessian at the global minimum point of the objective function. We propose a general framework that provides an efficient down-sampling strategy that can select data that preserves the strict positivity of the Hessian. Matrix sketching techniques from randomized linear algebra is heavily leaned on to achieve this goal. The method requires drawing samples from a certain distribution, and gradient free sampling methods are integrated to execute the data selection. Numerical experiments demonstrate the effectiveness of this method in selecting sensor locations for Schr\"odinger potential reconstruction.

Neural codecs have become crucial to recent speech and audio generation research. In addition to signal compression capabilities, discrete codecs have also been found to enhance downstream training efficiency and compatibility with autoregressive language models. However, as extensive downstream applications are investigated, challenges have arisen in ensuring fair comparisons across diverse applications. To address these issues, we present a new open-source platform ESPnet-Codec, which is built on ESPnet and focuses on neural codec training and evaluation. ESPnet-Codec offers various recipes in audio, music, and speech for training and evaluation using several widely adopted codec models. Together with ESPnet-Codec, we present VERSA, a standalone evaluation toolkit, which provides a comprehensive evaluation of codec performance over 20 audio evaluation metrics. Notably, we demonstrate that ESPnet-Codec can be integrated into six ESPnet tasks, supporting diverse applications., Comment: Accepted by SLT

LLMs' performance on complex tasks is still unsatisfactory. A key issue is that presently LLMs learn in a data-driven schema, while the instructions about these complex tasks are both scarce and hard to collect or construct. On the contrary, a prominent phenomenon is that LLMs can learn rather fast on simpler tasks with adequate prior knowledge captured during pretraining stage. Thus, if the prerequisite and mechanism of such rapid generalization could be elucidated, it could enhance the efficiency and effectiveness of the LLM's ability to learn complex tasks. Thus, in this paper, we employ a gradient-based method, to dissect the process that the SFT process adapts LLMs to downstream tasks via the perspective of attention patterns. We find that: (1) LLMs selectively activate task-specific attention heads during SFT; (2) activation patterns for complex tasks are combinations of basic task patterns; and (3) changes in a few parameters can significantly impact activation patterns after SFT on a small number of samples.Based on these insights, experiments are conducted to actually enhance the efficiency and effectiveness of SFT., Comment: in review

Large language models (LLMs) have shown outstanding performance across numerous real-world tasks. However, the autoregressive nature of these models makes the inference process slow and costly. Speculative decoding has emerged as a promising solution, leveraging a smaller auxiliary model to draft future tokens, which are then validated simultaneously by the larger model, achieving a speed-up of 1-2x. Although speculative decoding matches the same distribution as multinomial sampling, multinomial sampling itself is prone to suboptimal outputs, whereas beam sampling is widely recognized for producing higher-quality results by maintaining multiple candidate sequences at each step. This paper explores the novel integration of speculative decoding with beam sampling. However, there are four key challenges: (1) how to generate multiple sequences from the larger model's distribution given drafts sequences from the small model; (2) how to dynamically optimize the number of beams to balance efficiency and accuracy; (3) how to efficiently verify the multiple drafts in parallel; and (4) how to address the extra memory costs inherent in beam sampling. To address these challenges, we propose dynamic-width speculative beam decoding (DSBD). Specifically, we first introduce a novel draft and verification scheme that generates multiple sequences following the large model's distribution based on beam sampling trajectories from the small model. Then, we introduce an adaptive mechanism to dynamically tune the number of beams based on the context, optimizing efficiency and effectiveness. Besides, we extend tree-based parallel verification to handle multiple trees simultaneously, accelerating the verification process. Finally, we illustrate a simple modification to our algorithm to mitigate the memory overhead of beam sampling...

Silicon quantum dot devices stand as promising candidates for large-scale quantum computing due to their extended coherence times, compact size, and recent experimental demonstrations of sizable qubit arrays. Despite the great potential, controlling these arrays remains a significant challenge. This paper introduces a new virtual gate extraction method to quickly establish orthogonal control on the potentials for individual quantum dots. Leveraging insights from the device physics, the proposed approach significantly reduces the experimental overhead by focusing on crucial regions around charge state transition. Furthermore, by employing an efficient voltage sweeping method, we can efficiently pinpoint these charge state transition lines and filter out erroneous points. Experimental evaluation using real quantum dot chip datasets demonstrates a substantial 5.84x to 19.34x speedup over conventional methods, thereby showcasing promising prospects for accelerating the scaling of silicon spin qubit devices., Comment: 61st Design Automation Conference

The significant growth in the aquaculture industry over the last few decades encourages new technological and robotic solutions to help improve the efficiency and safety of production. In sea-based farming of Atlantic salmon in Norway, Unmanned Underwater Vehicles (UUVs) are already being used for inspection tasks. While new methods, systems and concepts for sub-sea operations are continuously being developed, these systems generally does not take into account how their presence might impact the fish. This abstract presents an experimental study on how underwater robotic operations at fish farms in Norway can affect farmed Atlantic salmon, and how the fish behaviour changes when exposed to the robot. The abstract provides an overview of the case study, the methods of analysis, and some preliminary results., Comment: 5 pages, 3 figures, accepted to IEEE IROS Workshop on Autonomous Robotic Systems in Aquaculture: Research Challenges and Industry Needs

Synthesizing camera movements from music and dance is highly challenging due to the contradicting requirements and complexities of dance cinematography. Unlike human movements, which are always continuous, dance camera movements involve both continuous sequences of variable lengths and sudden drastic changes to simulate the switching of multiple cameras. However, in previous works, every camera frame is equally treated and this causes jittering and unavoidable smoothing in post-processing. To solve these problems, we propose to integrate animator dance cinematography knowledge by formulating this task as a three-stage process: keyframe detection, keyframe synthesis, and tween function prediction. Following this formulation, we design a novel end-to-end dance camera synthesis framework \textbf{DanceCamAnimator}, which imitates human animation procedures and shows powerful keyframe-based controllability with variable lengths. Extensive experiments on the DCM dataset demonstrate that our method surpasses previous baselines quantitatively and qualitatively. Code will be available at \url{https://github.com/Carmenw1203/DanceCamAnimator-Official}., Comment: Accepted by ACM Multimedia 2024

Large language models (LLMs) with decoder-only architectures demonstrate remarkable in-context learning (ICL) capabilities. This feature enables them to effectively handle both familiar and novel tasks by utilizing examples provided within their input context. Recognizing the potential of this capability, we propose leveraging the ICL feature in LLMs to enhance the process of text embedding generation. To this end, we introduce a novel model bge-en-icl, which employs few-shot examples to produce high-quality text embeddings. Our approach integrates task-related examples directly into the query side, resulting in significant improvements across various tasks. Additionally, we have investigated how to effectively utilize LLMs as embedding models, including various attention mechanisms, pooling methods, etc. Our findings suggest that retaining the original framework often yields the best results, underscoring that simplicity is best. Experimental results on the MTEB and AIR-Bench benchmarks demonstrate that our approach sets new state-of-the-art (SOTA) performance. Our model, code and dataset are freely available at https://github.com/FlagOpen/FlagEmbedding .

Gamma-ray bursts (GRBs) are the most luminous transients in the universe. The interaction of the relativistic jet with the circumburst medium produces an afterglow and generates multiwavelength emission. In this work, we present simultaneous multiband photometry of GRB~240825A with the Multi-channel Photometric Survey Telescope (Mephisto) and analyze its temporal and spectral properties. The measurement began 128 seconds after the GRB trigger and continued until the fourth day when the afterglow essentially diminished and the measured brightness was close to that of the host galaxy. Based on the multiband light curves in the $uvgriz$ bands, we find that the optical flux density satisfies $F_{\nu,{\rm obs}}\propto t^{-1.34}\nu^{-2.48}$ with a spectral index of $2.48$ much larger than those of most other GRBs. To reconcile the measured much softer spectral energy distribution (SED) with that predicted by the standard afterglow model, an extra host-galaxy extinction of $E_{B-V}\sim(0.37-0.57)$ mag is required. We interpreted this excess as arising from a dense circumburst medium. We further find that the SED of the optical afterglow hardened as the afterglow decayed and the color excess $E_{B-V}$ decreased $\sim0.21$ mag in the first 3000 seconds. Finally, we analyze the properties of the host galaxy of GRB~240825A based on data from the SDSS, PanSTARRS and HSC-SSP surveys. For a host redshift of $z=0.659$, the stellar mass and star formation rate of the host galaxy are estimated to be $\log(M_*/M_\odot)=10.0^{+0.3}_{-0.3}$ and $\log({\rm SFR}/M_{\odot}{\rm yr}^{-1})= 0.6^{+0.8}_{-3.3}$, respectively, pointing to a gas-rich, star-forming, medium-size galaxy., Comment: 15 pages, 5 figures, 1 table. Comments welcome!

We present Persona-L, a novel approach for creating personas using Large Language Models (LLMs) and an ability-based framework, specifically designed to improve the representation of users with complex needs. Traditional methods of persona creation often fall short of accurately depicting the dynamic and diverse nature of complex needs, resulting in oversimplified or stereotypical profiles. Persona-L enables users to create and interact with personas through a chat interface. Persona-L was evaluated through interviews with UX designers (N=6), where we examined its effectiveness in reflecting the complexities of lived experiences of people with complex needs. We report our findings that indicate the potential of Persona-L to increase empathy and understanding of complex needs while also revealing the need for transparency of data used in persona creation, the role of the language and tone, and the need to provide a more balanced presentation of abilities with constraints.

Although current Multi-modal Large Language Models (MLLMs) demonstrate promising results in video understanding, processing extremely long videos remains an ongoing challenge. Typically, MLLMs struggle with handling thousands of visual tokens that exceed the maximum context length, and they suffer from the information decay due to token aggregation. Another challenge is the high computational cost stemming from the large number of video tokens. To tackle these issues, we propose Video-XL, an extra-long vision language model designed for efficient hour-scale video understanding. Specifically, we argue that LLMs can be adapted as effective visual condensers and propose Visual Context Latent Summarization which condenses visual contexts into highly compact forms. Extensive experiments demonstrate that our model achieves promising results on popular long video understanding benchmarks. For example, Video-XL outperforms the current state-of-the-art method on VNBench by nearly 10\% in accuracy. Moreover, Video-XL presents an impressive balance between efficiency and effectiveness, processing 2048 frames on a single 80GB GPU while achieving nearly 95% accuracy in the Needle-in-a-Haystack evaluation.

Magnetic Resonance Imaging (MRI) is a powerful, non-invasive diagnostic tool; however, its clinical applicability is constrained by prolonged acquisition times. Whilst present deep learning-based approaches have demonstrated potential in expediting MRI processes, these methods usually rely on known sampling patterns and exhibit limited generalisability to novel patterns. In the paper, we propose a sampling-pattern-agnostic MRI reconstruction method via a diffusion model through adaptive consistency enforcement. Our approach effectively reconstructs high-fidelity images with varied under-sampled acquisitions, generalising across contrasts and acceleration factors regardless of sampling trajectories. We train and validate across all contrasts in the MICCAI 2024 Cardiac MRI Reconstruction Challenge (CMRxRecon) dataset for the ``Random sampling CMR reconstruction'' task. Evaluation results indicate that our proposed method significantly outperforms baseline methods., Comment: MICCAI 2024 STACOM-CMRxRecon

During the fourth observing run, the LIGO-Virgo-KAGRA Collaboration reported the detection of a coalescing compact binary (GW230529$_{-}$181500) with component masses estimated at $2.5-4.5\, M_\odot$ and $1.2-2.0\, M_\odot$ with 90\% credibility. Given the current constraints on the maximum neutron star (NS) mass, this event is most likely a lower-mass-gap (LMG) black hole-neutron star (BHNS) binary. The spin magnitude of the BH, especially when aligned with the orbital angular momentum, is critical in determining whether the NS is tidally disrupted. An LMG BHNS merger with a rapidly spinning BH is an ideal candidate for producing electromagnetic counterparts. However, no such signals have been detected. In this study, we employ a detailed binary evolution model, incorporating new dynamical tide implementations, to explore the origin of BH spin in an LMG BHNS binary. If the NS forms first, the BH progenitor (He-rich star) must begin in orbit shorter than 0.35 days to spin up efficiently, potentially achieving a spin magnitude of $\chi_{\rm BH} > 0.3$. Alternatively, if a non-spinning BH (e.g., $M_{\rm BH} = 3.6\, M_\odot$) forms first, it can accrete up to $\sim 0.2\, M_\odot$ via Case BA mass transfer (MT), reaching a spin magnitude of $\chi_{\rm BH} \sim 0.18$ under Eddington-limited accretion. With a higher Eddington accretion limit (i.e., 10.0 $\Dot{M}_{\rm Edd}$), the BH can attain a significantly higher spin magnitude of $\chi_{\rm BH} \sim\,0.65$ by accreting approximately $1.0\, M_\odot$ during Case BA MT phase., Comment: Accepted by A&A Letters (09/11/2024)

HOPS 87 is a Class 0 protostellar core known to harbor an extremely young bipolar outflow and a hot corino. We report the discovery of localized, chemically rich regions near the bases of the two-lobe bipolar molecular outflow in HOPS 87 containing molecules such as H$_2$CO, $^{13}$CS, H$_2$S, OCS, and CH$_3$OH, the simplest complex organic molecule (COM). The locations and kinematics suggest that these localized features are due to jet-driven shocks rather than being part of the hot corino region encasing the protostar. The COM compositions of the molecular gas in these jet-localized regions are relatively simpler than those in the hot corino zone. We speculate that this simplicity is due to either the liberation of ice with a less complex chemical history or the effects of shock chemistry. Our study highlights the dynamic interplay between the protostellar bipolar outflow, disk, inner core environment, and the surrounding medium, contributing to our understanding of molecular complexity in solar-like young stellar objects., Comment: 16 pages, 6+2 figures, accepted by ApJ

Eye movement biometrics has received increasing attention thanks to its high secure identification. Although deep learning (DL) models have been recently successfully applied for eye movement recognition, the DL architecture still is determined by human prior knowledge. Differentiable Neural Architecture Search (DARTS) automates the manual process of architecture design with high search efficiency. DARTS, however, usually stacks the same multiple learned cells to form a final neural network for evaluation, limiting therefore the diversity of the network. Incidentally, DARTS usually searches the architecture in a shallow network while evaluating it in a deeper one, which results in a large gap between the architecture depths in the search and evaluation scenarios. To address this issue, we propose EM-DARTS, a hierarchical differentiable architecture search algorithm to automatically design the DL architecture for eye movement recognition. First, we define a supernet and propose a global and local alternate Neural Architecture Search method to search the optimal architecture alternately with an differentiable neural architecture search. The local search strategy aims to find an optimal architecture for different cells while the global search strategy is responsible for optimizing the architecture of the target network. To further reduce redundancy, a transfer entropy is proposed to compute the information amount of each layer, so as to further simplify search network. Our experiments on three public databases demonstrate that the proposed EM-DARTS is capable of producing an optimal architecture that leads to state-of-the-art recognition performance., Comment: Submited to IEEE Transactions on Information Forensics and Security

With the aid of large language models, current conversational recommender system (CRS) has gaining strong abilities to persuade users to accept recommended items. While these CRSs are highly persuasive, they can mislead users by incorporating incredible information in their explanations, ultimately damaging the long-term trust between users and the CRS. To address this, we propose a simple yet effective method, called PC-CRS, to enhance the credibility of CRS's explanations during persuasion. It guides the explanation generation through our proposed credibility-aware persuasive strategies and then gradually refines explanations via post-hoc self-reflection. Experimental results demonstrate the efficacy of PC-CRS in promoting persuasive and credible explanations. Further analysis reveals the reason behind current methods producing incredible explanations and the potential of credible explanations to improve recommendation accuracy., Comment: Findings of EMNLP 2024. Our code is available at https://github.com/mumen798/PC-CRS

This review paper, written for the second edition of the Handbook of Markov Chain Monte Carlo, provides an introduction to the study of convergence analysis for Markov chain Monte Carlo (MCMC), aimed at researchers new to the field. We focus on methods for constructing bounds on the distance between the distribution of a Markov chain at a given time and its stationary distribution. Two widely-used approaches are explored: the coupling method and the L2 theory of Markov chains. For the latter, we emphasize techniques based on conductance and isoperimetric inequalities. Additionally, we briefly discuss strategies for identifying slow convergence in Markov chains.

Transient surveys routinely detect supernovae (SNe) without obvious host galaxies. To understand the demographics of these "hostless" SNe and to constrain the possible host properties, we identify 161 SNe reported to the Transient Name Server since 2016 that do not have hosts cataloged from pre-explosion wide-field galaxy surveys. Using forced aperture photometry, we detect excess flux around only 56 of these SNe. Both thermonuclear and core-collapse (CC) SNe are present in our sample. Compared to flux-limited SNe samples with known hosts, superluminous supernovae (SLSNe), particularly hydrogen-deficient SLSNe, are over-represented here relative to all other SNe types; among CC SNe, there is also a higher fraction of interacting SNe than non-interacting. On the low-luminosity side, seven SNe have host absolute magnitude upper limits fainter than M_g=-12, about 1 per cent of the Small Magellanic Cloud's luminosity; the faintest limits are close to the luminosity of globular clusters or ultra-faint dwarf galaxies (M_g~-8). Fitting multi-band forced photometry, 11 SNe have host stellar masses <10^6 Msun assuming quiescent hosts, and 13 SNe have host stellar masses <10^5 Msun assuming star-forming hosts. The spatial distribution of hostless SNe indicates that the majority are not associated with known galaxy groups and clusters, ruling out intracluster stellar light as the primary contributor of such SNe. Hostless Type Ia SNe tend to be more luminous and slow-fading than SNe Ia with known host galaxies, implying a hidden population of low-mass and star-forming hosts. We conclude that any undetected host galaxies are likely star-forming dwarfs in the field., Comment: Accepted for publication in MNRAS

Fast radio burst (FRB) is a type of extragalactic radio signal characterized by millisecond duration, extremely high brightness temperature, and large dispersion measure. It remains a mystery in the universe. Advancements in instrumentation have led to the discovery of 816 FRB sources and 7622 bursts from 67 repeating FRBs now. This field is undergoing rapid development, rapidly advancing our understanding of the physics of FRBs as new observational data accumulates. The accumulation of data has also promoted our exploration of our universe. In this review, we summarize the statistical analysis and cosmological applications using large samples of FRBs, including the energy functions, the waiting time distributions of repeating FRBs, the probe of "missing baryons" and circumgalactic medium in the universe, measurements of cosmological parameters, exploration of the epoch of reionization history, and study of the gravitational lensing of FRBs., Comment: 29 pages, 12 figures. Currently under review. Comments welcome!

Sulfur-bearing molecules are commonly detected in dense cores within star-forming regions, but the total sulfur budget is significantly low, when compared to the interstellar medium (ISM) value. The properties of sulfur-bearing molecules are not well understood due to the absence of large sample studies with uniform observational configurations. To deepen our understanding of this subject, we conducted a study using ALMA 870 \micron~observations of 11 massive protoclusters. By checking the spectra of 248 dense cores in 11 massive protoclusters, a total of 10 sulfur-bearing species (CS, SO, \htcs, NS, \sot, \ttso, \tfsot, \ttsot, \seoo, \octfs) were identified. The parameters including systemic velocities, line widths, gas temperatures, column densities, and abundances were derived. Our results indicate that SO appears to be more easily detected in a wider range of physical environments than \htcs, despite these two species show similarities in gas distributions and abundances. \tfsot~and \htcs~are good tracers of the temperature of sulfur-bearing species, in which \htcs~traces the outer warm envelope and \tfsot~is associated with high-temperature central-regions. High-mass star-forming feedback (outflow and other non-thermal motions) significantly elevates the sulfur-bearing molecular abundances and detection rates specifically for \sot~and SO. A positive correlation between the \sot~abundance increasing factor ($F$) and temperatures suggests that \sot~could serve as a sulfur reservoir on the grain mantles of dense cores and then can be desorbed from dust to gas phase as the temperature rises. This work shows the importance of a large and unbiased survey to understand the sulfur depletion in dense cores.

The colossal parameters and computational overhead of Large Language Models (LLMs) challenge their real-world applications. Network pruning, which targets unstructured or structured sparsity by removing redundant parameters, has recently been explored for LLM acceleration. Existing LLM pruning works focus on unstructured pruning, which typically requires special hardware support for a practical speed-up. In contrast, structured pruning can reduce latency on general devices. However, it remains a challenge to perform structured pruning efficiently and maintain performance, especially at high sparsity ratios. To this end, we introduce an efficient structured pruning framework named CFSP, which leverages both Coarse (interblock) and Fine-grained (intrablock) activation information as an importance criterion to guide pruning. The pruning is highly efficient, as it only requires one forward pass to compute feature activations. Specifically, we first allocate the sparsity budget across blocks based on their importance and then retain important weights within each block. In addition, we introduce a recovery fine-tuning strategy that adaptively allocates training overhead based on coarse-grained importance to further improve performance. Experimental results demonstrate that CFSP outperforms existing methods on diverse models across various sparsity budgets. Our code will be available at https://github.com/wyxscir/CFSP., Comment: Work in progress

We propose an adaptive fine-tuning algorithm for multimodal large models. The core steps of this algorithm involve two stages of truncation. First, the vast amount of data is projected into a semantic vector space, and the MiniBatchKMeans algorithm is used for automated clustering. This classification ensures that the data within each cluster exhibit high semantic similarity. Next, we process the data in each cluster, calculating the translational difference between the original and perturbed data in the multimodal large model's vector space. This difference serves as a generalization metric for the data. Based on this metric, we select the data with high generalization potential for training. We applied this algorithm to train the InternLM-XComposer2-VL-7B model on two 3090 GPUs using one-third of the GeoChat multimodal remote sensing dataset. The results demonstrate that our algorithm outperforms the state-of-the-art baselines. various baselines. The model trained on our optimally chosen one-third dataset, based on experimental validation, exhibited only 1% reduction in performance across various remote sensing metrics compared to the model trained on the full dataset. This approach significantly preserved general-purpose capabilities while reducing training time by 68.2%. Furthermore, the model achieved scores of 89.86 and 77.19 on the UCMerced and AID evaluation datasets, respectively, surpassing the GeoChat dataset by 5.43 and 5.16 points. It only showed a 0.91-point average decrease on the LRBEN evaluation dataset.

We present high-spatial-resolution (0.7 to 1.0 arcsec) submillimeter observations of continuum and molecular lines of CH3OCHO, CH3OCH3, and H2CCO toward 11 high-mass star-forming regions using the Atacama Large Millimetre/submillimetre Array (ALMA). A total of 19 separate cores from 9 high-mass star-forming regions are found to be line-rich, including high-, intermediate-, and low-mass line-rich cores. The three molecules are detected in these line-rich cores. We map the emission of CH3OCHO, CH3OCH3, and H2CCO in 9 high-mass star-forming regions. The spatial distribution of the three molecules is very similar and concentrated in the areas of intense continuum emission. We also calculate the rotation temperatures, column densities, and abundances of CH3OCHO, CH3OCH3, and H2CCO under the local thermodynamic equilibrium (LTE) assumption. The abundances relative to H2 and CH3OH, and line widths of the three molecules are significantly correlated. The abundances relative to H2, temperatures and line widths of the three molecules tend to be higher in cores with higher mass and outflows detected. The possible chemical links of the three molecules are discussed., Comment: 35 pages, 9 figures

Reward models (RMs) play a pivotal role in aligning large language models (LLMs) with human preferences. However, traditional RM training, which relies on response pairs tied to specific prompts, struggles to disentangle prompt-driven preferences from prompt-independent artifacts, such as response length and format. In this work, we expose a fundamental limitation of current RM training methods, where RMs fail to effectively distinguish between contextual signals and irrelevant artifacts when determining preferences. To address this, we introduce a causal framework that learns preferences independent of these artifacts and propose a novel data augmentation technique designed to eliminate them. Extensive experiments show that our approach successfully filters out undesirable artifacts, yielding a more robust reward model (RRM). Our RRM improves the performance of a pairwise reward model trained on Gemma-2-9b-it, on RewardBench, increasing accuracy from 80.61% to 84.15%. Additionally, we train two DPO policies using both the RM and RRM, demonstrating that the RRM significantly enhances DPO-aligned policies, improving MT-Bench scores from 7.27 to 8.31 and length-controlled win-rates in AlpacaEval-2 from 33.46% to 52.49%.

As text-based speech editing becomes increasingly prevalent, the demand for unrestricted free-text editing continues to grow. However, existing speech editing techniques encounter significant challenges, particularly in maintaining intelligibility and acoustic consistency when dealing with out-of-domain (OOD) text. In this paper, we introduce, DiffEditor, a novel speech editing model designed to enhance performance in OOD text scenarios through semantic enrichment and acoustic consistency. To improve the intelligibility of the edited speech, we enrich the semantic information of phoneme embeddings by integrating word embeddings extracted from a pretrained language model. Furthermore, we emphasize that interframe smoothing properties are critical for modeling acoustic consistency, and thus we propose a first-order loss function to promote smoother transitions at editing boundaries and enhance the overall fluency of the edited speech. Experimental results demonstrate that our model achieves state-of-the-art performance in both in-domain and OOD text scenarios.

Heavy flavor jets provide ideal tools to probe the mass effect on jet substructure in both vacuum and quark-gluon plasma (QGP). Energy-energy correlator (EEC) is an excellent jet substructure observable owning to its strong sensitivity to jet physics at different scales. We perform a complete realistic simulation on medium modification of heavy and light flavor jet EEC in heavy-ion collisions. A clear flavor hierarchy is observed for jet EEC in both vacuum and QGP due to the mass effect. The medium modification of inclusive jet EEC at different angular scales exhibits very rich structure: suppression at intermediate angles, and enhancement at small and large angles, which can be well explained by the interplay of mass effect, energy loss, medium-induced radiation and medium response. These unique features of jet EEC are shown to probe the physics of jet-medium interaction at different scales, and can be readily validated by upcoming experiments., Comment: 7 pages, 5 figures

With the recent evolution of large language models (LLMs), concerns about aligning such models with human values have grown. Previous research has primarily focused on assessing LLMs' performance in terms of the Helpful, Honest, Harmless (3H) basic principles, while often overlooking their alignment with educational values in the Chinese context. To fill this gap, we present Edu-Values, the first Chinese education values evaluation benchmark designed to measure LLMs' alignment ability across seven dimensions: professional ideology, cultural literacy, educational knowledge and skills, education laws and regulations, teachers' professional ethics, basic competencies, and subject knowledge. We meticulously design and compile 1,418 questions, including multiple-choice, multi-modal question answering, subjective analysis, adversarial prompts, and questions on traditional Chinese culture. We conduct both human evaluation and automatic evaluation over 11 state-of-the-art (SoTA) LLMs, and highlight three main findings: (1) due to differences in educational culture, Chinese LLMs significantly outperform English LLMs, with Qwen 2 ranking the first with a score of 81.37; (2) LLMs perform well in subject knowledge and teaching skills but struggle with teachers' professional ethics and basic competencies; (3) LLMs excel at multiple-choice questions but perform poorly on subjective analysis and multi-modal tasks. This demonstrates the effectiveness and potential of the proposed benchmark. Our dataset is available at https://github.com/zhangpeii/Edu-Values.git., Comment: 20 pages, 7 figures

The differential branching fraction and angular coefficients of \ensuremath{\it{\Lambda}^\mathrm{0}_b \rightarrow pK^-\mu^+\mu^-}\xspace decays are measured in bins of the dimuon mass squared and dihadron mass. The analysis is performed using a data set corresponding to 9$\aunit{fb}^{-1}$ of integrated luminosity collected with the $\mbox{LHCb}$ detector between 2011 and 2018. The data are consistent with receiving contributions from a mixture of $\it{\Lambda}$ resonances with different spin-parity quantum numbers. The angular coefficients show a pattern of vector--axial vector interference that is a characteristic of the type of flavour-changing neutral-current transition relevant for these decays., Comment: All figures and tables, along with machine-readable versions and any supplementary material and additional information, are available at https://cern.ch/lhcbproject/Publications/p/3264.html (LHCb public pages)

Quantum state tomography (QST) via local measurements on reduced density matrices (LQST) is a promising approach but becomes impractical for large systems. To tackle this challenge, we developed an efficient quantum state tomography method inspired by quantum overlapping tomography [Phys. Rev. Lett. 124, 100401(2020)], which utilizes parallel measurements (PQST). In contrast to LQST, PQST significantly reduces the number of measurements and offers more robustness against shot noise. Experimentally, we demonstrate the feasibility of PQST in a tree-like superconducting qubit chip by designing high-efficiency circuits, preparing W states, ground states of Hamiltonians and random states, and then reconstructing these density matrices using full quantum state tomography (FQST), LQST, and PQST. Our results show that PQST reduces measurement cost, achieving fidelities of 98.68\% and 95.07\% after measuring 75 and 99 observables for 6-qubit and 9-qubit W states, respectively. Furthermore, the reconstruction of the largest density matrix of the 12-qubit W state is achieved with the similarity of 89.23\% after just measuring $243$ parallel observables, while $3^{12}=531441$ complete observables are needed for FQST. Consequently, PQST will be a useful tool for future tasks such as the reconstruction, characterization, benchmarking, and properties learning of states., Comment: To appear in PRL(2024)

2D metal halide perovskites have enabled significant stability improvements in perovskite devices, particularly in resistance to moisture. However, some 2D perovskites are even more susceptible to photooxidation compared to 3D perovskites. This is particularly true for more commonly investigated Ruddlesden-Popper (RP) perovskites that exhibit increased susceptibility to photoinduced degradation compared to Dion-Jacobson (DJ) perovskites. Comparisons between different RP and DJ perovskites reveal that this phenomenon cannot be explained by commonly proposed differences in superoxide ion generation, interlayer distance and lattice structural rigidity differences. Instead, the resistance to photooxidation of DJ perovskites can be attributed to decreased likelihood of double deprotonation events (compared to single deprotonation events in RP perovskites) required for the loss of organic cations and the perovskite decomposition. Consequently, DJ perovskites are less susceptible to oxidative degradation (both photo- and electrochemically induced), which leads to improved operational stability of solar cells based on these materials., Comment: Main text: 19 pages, 6 figures, supplementary information: 62 pages, 47 figures

Video generation using diffusion-based models is constrained by high computational costs due to the frame-wise iterative diffusion process. This work presents a Diffusion Reuse MOtion (Dr. Mo) network to accelerate latent video generation. Our key discovery is that coarse-grained noises in earlier denoising steps have demonstrated high motion consistency across consecutive video frames. Following this observation, Dr. Mo propagates those coarse-grained noises onto the next frame by incorporating carefully designed, lightweight inter-frame motions, eliminating massive computational redundancy in frame-wise diffusion models. The more sensitive and fine-grained noises are still acquired via later denoising steps, which can be essential to retain visual qualities. As such, deciding which intermediate steps should switch from motion-based propagations to denoising can be a crucial problem and a key tradeoff between efficiency and quality. Dr. Mo employs a meta-network named Denoising Step Selector (DSS) to dynamically determine desirable intermediate steps across video frames. Extensive evaluations on video generation and editing tasks have shown that Dr. Mo can substantially accelerate diffusion models in video tasks with improved visual qualities.

Retrieval Augmented Generation (RAG) improves large language models (LMs) by incorporating non-parametric knowledge through evidence retrieval from external sources. However, it often struggles to filter out inconsistent and irrelevant information that can distract the LM from its tasks. While compressing the retrieved evidence with a compression model aims to address this issue, the compressed evidence may still be unfamiliar to the target model used for downstream task, potentially failing to utilize the evidence effectively. We propose FaviComp (Familiarity-aware Evidence Compression), a novel training-free evidence compression technique that makes retrieved evidence more familiar to the target model, while seamlessly integrating parametric knowledge from the model. Specifically, FaviComp proactively lowers the perplexity of the compressed evidence with regard to the target model by combining token probabilities from both the compression model and the target model to generate context that is more familiar to the target model. This approach balances the integration of parametric and non-parametric knowledge, which is especially helpful in complex tasks where the retrieved evidence set may not contain all the necessary information. Experimental results demonstrate that FaviComp consistently outperforms existing baselines in multiple open-domain QA datasets, achieving high compression rates and showcasing the effective integration of both parametric and non-parametric knowledge.

Diffusion-based text-to-audio (TTA) generation has made substantial progress, leveraging latent diffusion model (LDM) to produce high-quality, diverse and instruction-relevant audios. However, beyond generation, the task of audio editing remains equally important but has received comparatively little attention. Audio editing tasks face two primary challenges: executing precise edits and preserving the unedited sections. While workflows based on LDMs have effectively addressed these challenges in the field of image processing, similar approaches have been scarcely applied to audio editing. In this paper, we introduce AudioEditor, a training-free audio editing framework built on the pretrained diffusion-based TTA model. AudioEditor incorporates Null-text Inversion and EOT-suppression methods, enabling the model to preserve original audio features while executing accurate edits. Comprehensive objective and subjective experiments validate the effectiveness of AudioEditor in delivering high-quality audio edits. Code and demo can be found at https://github.com/NKU-HLT/AudioEditor.

With the advent of high-precision nanoscale lithography technology, high-resolution image sensing has experienced rapid development in recent years. Currently, mainstream commercial image sensors predominantly utilize Bayer array color filters to implement RGB colorful imaging strategies. However, as pixel sizes transition into the submicron dimensions, traditional dye filters used in image sensors have long been hampered by limited optical efficiency, suboptimal signal-to-noise ratios, and significant difficulties in miniaturization. In this work, a novel 4-channel RGB-IR color router for image sensing, distinct from the traditional absorption-transmission mechanisms, was proposed through inverse design methodologies. Utilizing genetic algorithms and DCGAN models, approximately 20,000 random color routing structures were generated and trained. From these, an optimized spectral splitting structure with a minimal periodic size of 1.6 um * 1.6 um was identified. This structure achieves peak optical efficiencies 1.7 times greater than those of dye filters, while also offering superior color imaging quality and signal intensity. This innovative design approach, leveraging deep learning integration, demonstrates an on-chip strategy for color realization in 4-channel image sensors, and holds significant promise for enhancing the development of next-generation high-performance image sensing chip systems., Comment: 19 pages,5 figures

Large Language Models (LLMs) have demonstrated remarkable performance through supervised fine-tuning or in-context learning using gold labels. However, this paradigm is limited by the availability of gold labels, while in certain scenarios, LLMs may need to perform tasks that are too complex for humans to provide such labels. To tackle this challenge, this study explores whether solely utilizing unlabeled data can elicit strong model capabilities. We propose a new paradigm termed zero-to-strong generalization. We iteratively prompt LLMs to annotate unlabeled data and retain high-quality labels by filtering. Surprisingly, we obverse that this iterative process gradually unlocks LLMs' potential on downstream tasks. Our experiments on extensive classification and reasoning tasks confirm the effectiveness of our proposed framework. Our analysis indicates that this paradigm is effective for both in-context learning and fine-tuning, and for various model sizes., Comment: 15 pages