Search

The integrated sensing and communication (ISAC) system under multi-input multi-output (MIMO) architecture achieves dual functionalities of sensing and communication on the same platform by utilizing spatial gain, which provides a feasible paradigm facing spectrum congestion. However, the dual functionalities of sensing and communication operating simultaneously in the same platform bring severe interference in the ISAC systems. Facing this challenge, we propose a joint optimization framework for transmit beamforming and receive filter design for ISAC systems with MIMO architecture. We aim to maximize the signal-to-clutter-plus-noise ratio (SCNR) at the receiver while considering various constraints such as waveform similarity, power budget, and communication performance requirements to ensure the integration of the dual functionalities. In particular, the overall transmit beamforming is refined into sensing beamforming and communication beamforming, and a quadratic transformation (QT) is introduced to relax and convert the complex non-convex optimization objective. An efficient algorithm based on covariance matrix tapers (CMT) is proposed to restructure the clutter covariance matrix considering the mismatched steering vector, thereby improving the robustness of the ISAC transceiver design. Numerical simulations are provided to demonstrate the effectiveness of the proposed algorithm.

Underwater image enhancement (UIE) has attracted much attention owing to its importance for underwater operation and marine engineering. Motivated by the recent advance in generative models, we propose a novel UIE method based on image-conditional diffusion transformer (ICDT). Our method takes the degraded underwater image as the conditional input and converts it into latent space where ICDT is applied. ICDT replaces the conventional U-Net backbone in a denoising diffusion probabilistic model (DDPM) with a transformer, and thus inherits favorable properties such as scalability from transformers. Furthermore, we train ICDT with a hybrid loss function involving variances to achieve better log-likelihoods, which meanwhile significantly accelerates the sampling process. We experimentally assess the scalability of ICDTs and compare with prior works in UIE on the Underwater ImageNet dataset. Besides good scaling properties, our largest model, ICDT-XL/2, outperforms all comparison methods, achieving state-of-the-art (SOTA) quality of image enhancement.

Large language models~(LLMs) have demonstrated impressive performance in various applications, among which role-playing language agents (RPLAs) have engaged a broad user base. Now, there is a growing demand for RPLAs that represent Key Opinion Leaders (KOLs), \ie, Internet celebrities who shape the trends and opinions in their domains. However, research in this line remains underexplored. In this paper, we hence introduce MINDECHO, a comprehensive framework for the development and evaluation of KOL RPLAs. MINDECHO collects KOL data from Internet video transcripts in various professional fields, and synthesizes their conversations leveraging GPT-4. Then, the conversations and the transcripts are used for individualized model training and inference-time retrieval, respectively. Our evaluation covers both general dimensions (\ie, knowledge and tones) and fan-centric dimensions for KOLs. Extensive experiments validate the effectiveness of MINDECHO in developing and evaluating KOL RPLAs.

Visual tracking has advanced significantly in recent years, mainly due to the availability of large-scale training datasets. These datasets have enabled the development of numerous algorithms that can track objects with high accuracy and robustness.However, the majority of current research has been directed towards tracking generic objects, with less emphasis on more specialized and challenging scenarios. One such challenging scenario involves tracking reflected objects. Reflections can significantly distort the appearance of objects, creating ambiguous visual cues that complicate the tracking process. This issue is particularly pertinent in applications such as autonomous driving, security, smart homes, and industrial production, where accurately tracking objects reflected in surfaces like mirrors or glass is crucial. To address this gap, we introduce TRO, a benchmark specifically for Tracking Reflected Objects. TRO includes 200 sequences with around 70,000 frames, each carefully annotated with bounding boxes. This dataset aims to encourage the development of new, accurate methods for tracking reflected objects, which present unique challenges not sufficiently covered by existing benchmarks. We evaluated 20 state-of-the-art trackers and found that they struggle with the complexities of reflections. To provide a stronger baseline, we propose a new tracker, HiP-HaTrack, which uses hierarchical features to improve performance, significantly outperforming existing algorithms. We believe our benchmark, evaluation, and HiP-HaTrack will inspire further research and applications in tracking reflected objects. The TRO and code are available at https://github.com/OpenCodeGithub/HIP-HaTrack.

Large-deformation localized failure in geomaterials plays a crucial role in geohazards engineering. This article investigates shear bands and retrogressive failure of geomaterials through a recently formulated large-deformation-plasticity periporomechanics (PPM) paradigm. Periporomechanics is a nonlocal reformulation of classical poromechanics through the effective force and peridynamic state concepts. The nonlocal deformation gradient is multiplicatively decomposed into the elastic and plastic parts in this large-deformation PPM paradigm. The stabilized correspondence principle is adopted to implement a classical elastoplastic constitutive model into the new PPM paradigm. We have numerically implemented this large-deformation plasticity PPM paradigm through a Lagrangian meshfree method in space and an explicit Newmark scheme in time. The implemented PPM framework is used to model shear banding and retrogressive slope failure in porous media under dry conditions. The numerical results have demonstrated the efficacy and robustness of this new PPM paradigm for modeling shear banding and retrogressive failure involving large deformation in porous media.

Performance plays a critical role in ensuring the smooth operation of any mobile application, directly influencing user engagement and retention. Android applications are no exception. However, unlike functionality issues, performance issues are more challenging to discover as their root causes are sophisticated and typically emerge under specific payloads. To tackle this problem, researchers have dedicated substantial efforts to proposing automatic approaches for understanding, detecting, and resolving performance issues. Despite these endeavors, it still remains unknown what the status quo of Android performance analysis is, and whether existing approaches can indeed accurately reflect real performance issues. To fill this research gap, we conducted a systematic literature review followed by an explanatory study to explore relevant studies and real-world challenges. Our findings reveal that current tools have limited capabilities, covering only 17.50% of the performance issues. Additionally, existing datasets encompass only 27.50% of the issues and are very limited in size. We also show real-world issue patterns, underscoring the huge gap between the identified techniques and practical concerns. Furthermore, possible solutions are provided to guide future research towards achieving effective performance issue detection and resolution.

The next sixth generation (6G) networks are envisioned to integrate sensing and communications in a single system, thus greatly improving spectrum utilization and reducing hardware costs. Low earth orbit (LEO) satellite communications combined with massive multiple-input multiple-output (MIMO) technology holds significant promise in offering ubiquitous and seamless connectivity with high data rates. Existing integrated sensing and communications (ISAC) studies mainly focus on terrestrial systems, while operating ISAC in massive MIMO LEO satellite systems is promising to provide high-capacity communication and flexible sensing ubiquitously. In this paper, we first give an overview of LEO satellite systems and ISAC and consider adopting ISAC in the massive MIMO LEO satellite systems. Then, the recent research advances are presented. A discussion on related challenges and key enabling technologies follows. Finally, we point out some open issues and promising research directions., Comment: 6 pages,4 figures

We develop a new framework to prove the mixing or relaxation time for the Glauber dynamics on spin systems with unbounded degree. It works for general spin systems including both $2$-spin and multi-spin systems. As applications for this approach: $\bullet$ We prove the optimal $O(n)$ relaxation time for the Glauber dynamics of random $q$-list-coloring on an $n$-vertices triangle-tree graph with maximum degree $\Delta$ such that $q/\Delta > \alpha^\star$, where $\alpha^\star \approx 1.763$ is the unique positive solution of the equation $\alpha = \exp(1/\alpha)$. This improves the $n^{1+o(1)}$ relaxation time for Glauber dynamics obtained by the previous work of Jain, Pham, and Vuong (2022). Besides, our framework can also give a near-linear time sampling algorithm under the same condition. $\bullet$ We prove the optimal $O(n)$ relaxation time and near-optimal $\widetilde{O}(n)$ mixing time for the Glauber dynamics on hardcore models with parameter $\lambda$ in $\textit{balanced}$ bipartite graphs such that $\lambda < \lambda_c(\Delta_L)$ for the max degree $\Delta_L$ in left part and the max degree $\Delta_R$ of right part satisfies $\Delta_R = O(\Delta_L)$. This improves the previous result by Chen, Liu, and Yin (2023). At the heart of our proof is the notion of $\textit{coupling independence}$ which allows us to consider multiple vertices as a huge single vertex with exponentially large domain and do a "coarse-grained" local-to-global argument on spin systems. The technique works for general (multi) spin systems and helps us obtain some new comparison results for Glauber dynamics.

We address the convergence rate of Markov chains for randomly generating an edge coloring of a given tree. Our focus is on the Glauber dynamics which updates the color at a randomly chosen edge in each step. For a tree $T$ with $n$ vertices and maximum degree $\Delta$, when the number of colors $q$ satisfies $q\geq\Delta+2$ then we prove that the Glauber dynamics has an optimal relaxation time of $O(n)$, where the relaxation time is the inverse of the spectral gap. This is optimal in the range of $q$ in terms of $\Delta$ as Dyer, Goldberg, and Jerrum (2006) showed that the relaxation time is $\Omega(n^3)$ when $q=\Delta+1$. For the case $q=\Delta+1$, we show that an alternative Markov chain which updates a pair of neighboring edges has relaxation time $O(n)$. Moreover, for the $\Delta$-regular complete tree we prove $O(n\log^2{n})$ mixing time bounds for the respective Markov chain. Our proofs establish approximate tensorization of variance via a novel inductive approach, where the base case is a tree of height $\ell=O(\Delta^2\log^2{\Delta})$, which we analyze using a canonical paths argument.

The reward model has become increasingly important in alignment, assessment, and data construction for large language models (LLMs). Most existing researchers focus on enhancing reward models through data improvements, following the conventional training framework for reward models that directly optimizes the predicted rewards. In this paper, we propose a hybrid alignment framework HaF-RM for reward model training by introducing an additional constraint on token-level policy probabilities in addition to the reward score. It can simultaneously supervise the internal preference model at the token level and optimize the mapping layer of the reward model at the sequence level. Theoretical justifications and experiment results on five datasets show the validity and effectiveness of our proposed hybrid framework for training a high-quality reward model. By decoupling the reward modeling procedure and incorporating hybrid supervision, our HaF-RM framework offers a principled and effective approach to enhancing the performance and alignment of reward models, a critical component in the responsible development of powerful language models. We release our code at https://haf-rm.github.io.

Metaphors are part of everyday language and shape the way in which we conceptualize the world. Moreover, they play a multifaceted role in communication, making their understanding and generation a challenging task for language models (LMs). While there has been extensive work in the literature linking metaphor to the fulfilment of individual intentions, no comprehensive taxonomy of such intentions, suitable for natural language processing (NLP) applications, is available to present day. In this paper, we propose a novel taxonomy of intentions commonly attributed to metaphor, which comprises 9 categories. We also release the first dataset annotated for intentions behind metaphor use. Finally, we use this dataset to test the capability of large language models (LLMs) in inferring the intentions behind metaphor use, in zero- and in-context few-shot settings. Our experiments show that this is still a challenge for LLMs.

Objects in aerial images are typically embedded in complex backgrounds and exhibit arbitrary orientations. When employing oriented bounding boxes (OBB) to represent arbitrary oriented objects, the periodicity of angles could lead to discontinuities in label regression values at the boundaries, inducing abrupt fluctuations in the loss function. To address this problem, an OBB representation based on the complex plane is introduced in the oriented detection framework, and a trigonometric loss function is proposed. Moreover, leveraging prior knowledge of complex background environments and significant differences in large objects in aerial images, a conformer RPN head is constructed to predict angle information. The proposed loss function and conformer RPN head jointly generate high-quality oriented proposals. A category-aware dynamic label assignment based on predicted category feedback is proposed to address the limitations of solely relying on IoU for proposal label assignment. This method makes negative sample selection more representative, ensuring consistency between classification and regression features. Experiments were conducted on four realistic oriented detection datasets, and the results demonstrate superior performance in oriented object detection with minimal parameter tuning and time costs. Specifically, mean average precision (mAP) scores of 82.02%, 71.99%, 69.87%, and 98.77% were achieved on the DOTA-v1.0, DOTA-v1.5, DIOR-R, and HRSC2016 datasets, respectively.

With the emergence of the Software 3.0 era, there is a growing trend of compressing and integrating large models into software systems, with significant societal implications. Regrettably, in numerous instances, model compression techniques impact the fairness performance of these models and thus the ethical behavior of DNN-powered software. One of the most notable example is the Lottery Ticket Hypothesis (LTH), a prevailing model pruning approach. This paper demonstrates that fairness issue of LTHbased pruning arises from both its subnetwork selection and training procedures, highlighting the inadequacy of existing remedies. To address this, we propose a novel pruning framework, Ballot, which employs a novel conflict-detection-based subnetwork selection to find accurate and fair subnetworks, coupled with a refined training process to attain a high-performance model, thereby improving the fairness of DNN-powered software. By means of this procedure, Ballot improves the fairness of pruning by 38.00%, 33.91%, 17.96%, and 35.82% compared to state-of-the-art baselines, namely Magnitude Pruning, Standard LTH, SafeCompress, and FairScratch respectively, based on our evaluation of five popular datasets and three widely used models. Our code is available at https://anonymous.4open.science/r/Ballot-506E.

The detection and classification of power quality disturbances (PQDs) carries significant importance for power systems. In response to this imperative, numerous intelligent diagnostic methods have been developed. However, existing identification methods usually concentrate on single-type signals or on complex signals with two types, rendering them susceptible to noisy labels and environmental effects. This study proposes a novel method for the classification of PQDs, termed ST-GSResNet, which utilizes the S-Transform and an improved residual neural network (ResNet) with a channel attention mechanism. The ST-GSResNet approach initially uses the S-Transform to transform a time-series signal into a 2D time-frequency image for feature enhancement. Then, an improved ResNet model is introduced, which employs grouped convolution instead of the traditional convolution operation. This improvement aims to facilitate learning with a block-diagonal structured sparsity on the channel dimension, the highly-correlated filters are learned in a more structured way in the networks with filter groups. By reducing the number of parameters in the network in this significant manner, the model becomes less prone to overfitting. Furthermore, the SE module concentrates on primary components, which enhances the model's robustness in recognition and immunity to noise. Experimental results demonstrate that, compared to existing deep learning models, our approach has advantages in computational efficiency and classification accuracy.

The Multi-Output Gaussian Process is is a popular tool for modelling data from multiple sources. A typical choice to build a covariance function for a MOGP is the Linear Model of Coregionalization (LMC) which parametrically models the covariance between outputs. The Latent Variable MOGP (LV-MOGP) generalises this idea by modelling the covariance between outputs using a kernel applied to latent variables, one per output, leading to a flexible MOGP model that allows efficient generalization to new outputs with few data points. Computational complexity in LV-MOGP grows linearly with the number of outputs, which makes it unsuitable for problems with a large number of outputs. In this paper, we propose a stochastic variational inference approach for the LV-MOGP that allows mini-batches for both inputs and outputs, making computational complexity per training iteration independent of the number of outputs., Comment: none

This paper considers the problem of scheduling uplinks and downlinks transmissions in an Internet of Things (IoT) network that uses a mode-based time structure and Rate Splitting Multiple Access (RSMA). Further, devices employ power splitting to harvest energy and receive data simultaneously from a Hybrid Access Point (HAP). To this end, this paper outlines a Mixed Integer Linear Program (MILP) that can be employed by a HAP to optimize the following quantities over a given time horizon: (i) mode (downlink or uplink) of time slots, (ii) transmit power of each packet, (iii) power splitting ratio of devices, and (iv) decoding order in uplink slots. The MILP yields the optimal number of packet transmissions over a given planning horizon given non-causal channel state information. We also present a learning based approach to determine the mode of each time slot using causal channel state information. The results show that the learning based approach achieves 90% of the optimal number of packet transmissions, and the HAP receives 25% more packets as compared to competing approaches., Comment: 12 pages, 7 figures

Using $(27.12\pm 0.14)\times10^8$ $\psi(3686)$ decays and data samples of $e^+e^-$ collisions with $\sqrt{s}$ from 4.130 to 4.780~GeV collected with the BESIII detector, we report the first observation of the electromagnetic Dalitz transition $h_c\to e^+e^-\eta_c$ with a statistical significance of $5.4\sigma$. We measure the ratio of the branching fractions $\frac{\mathcal{B}(h_c\rightarrow e^+e^-\eta_c)}{\mathcal{B}(h_c\rightarrow \gamma \eta_c)}$ separately for the $h_c$ samples produced via $\psi(3686)\to\pi^0h_c$ and $e^+e^-\to\pi^+\pi^-h_c$. The average ratio is determined to be $(0.59\pm0.10(\text{stat.})\pm0.04(\text{syst.}))\%$, where the uncertainty includes both statistical and systematic components.

Bilevel optimization has shown its utility across various machine learning settings, yet most algorithms in practice require second-order information, making it challenging to scale them up. Only recently, a paradigm of first-order algorithms emerged, capable of effectively addressing bilevel optimization problems. Nevertheless, the practical efficiency of this paradigm remains unverified, particularly in the context of large language models (LLMs). This paper introduces the first scalable instantiation of this paradigm called ScaleBiO, focusing on bilevel optimization for large-scale LLM data reweighting. By combining with a recently proposed memory-efficient training technique called LISA, our novel algorithm allows the paradigm to scale to 34-billion-parameter LLMs on eight A40 GPUs, marking the first successful application of bilevel optimization under practical scenarios for large-sized LLMs. Empirically, extensive experiments on data reweighting verify the effectiveness of ScaleBiO for different-scaled models, including GPT-2, LLaMA-3-8B, GPT-NeoX-20B, and Yi-34B, where bilevel optimization succeeds in filtering irrelevant data samples and selecting informative samples. Theoretically, ScaleBiO ensures the optimality of the learned data weights, along with a convergence guarantee matching the conventional first-order bilevel optimization paradigm on smooth and strongly convex objectives.

Multipartite entanglement is an essential resource for quantum information theory and technologies, but its quantification has been a persistent challenge. Recently, Concentratable Entanglement (CE) has been introduced as a promising candidate for a multipartite entanglement measure, which can be efficiently estimated across two state copies. In this work, we introduce Generalized Concentratable Entanglement (GCE) measures, highlight a natural correspondence to quantum Tsallis entropies, and conjecture a new entropic inequality that may be of independent interest. We show how to efficiently measure the GCE in a quantum computer, using parallelized permutation tests across a prime number of state copies. We exemplify the practicality of such computation for probabilistic entanglement concentration into W states with three state copies. Moreover, we show that an increased number of state copies provides an improved error bound on this family of multipartite entanglement measures in the presence of imperfections. Finally, we prove that GCE is still a well-defined entanglement monotone as its value, on average, does not increase under local operations and classical communication (LOCC)., Comment: 5 pages + 17 pages appendix; 4 + 2 figures; 1 table

We mainly investigate the continuous dependence on parameters of nontrivial solutions for a generalized poly-Laplacian system on the weighted finite graph $G=(V, E)$. We firstly present an existence result of mountain pass type nontrivial solutions when the nonlinear term $F$ satisfies the super-$(p, q)$ linear growth condition which is a simple generalization of those results in [28]. Then we mainly show that the mountain pass type nontrivial solutions of the poly-Laplacian system are uniformly bounded for parameters and the concrete upper and lower bounds are given, and are continuously dependent on parameters. Similarly, we also present the existence result, the concrete upper and lower bounds, uniqueness, and dependence on parameters for the locally minimum type nontrivial solutions. Subsequently, we present an example on optimal control as an application of our results. Finally, we give a nonexistence result and some results for the corresponding scalar equation.

As LLMs rapidly advance, increasing concerns arise regarding risks about actual authorship of texts we see online and in real world. The task of distinguishing LLM-authored texts is complicated by the nuanced and overlapping behaviors of both machines and humans. In this paper, we challenge the current practice of considering LLM-generated text detection a binary classification task of differentiating human from AI. Instead, we introduce a novel ternary text classification scheme, adding an "undecided" category for texts that could be attributed to either source, and we show that this new category is crucial to understand how to make the detection result more explainable to lay users. This research shifts the paradigm from merely classifying to explaining machine-generated texts, emphasizing need for detectors to provide clear and understandable explanations to users. Our study involves creating four new datasets comprised of texts from various LLMs and human authors. Based on new datasets, we performed binary classification tests to ascertain the most effective SOTA detection methods and identified SOTA LLMs capable of producing harder-to-detect texts. We constructed a new dataset of texts generated by two top-performing LLMs and human authors, and asked three human annotators to produce ternary labels with explanation notes. This dataset was used to investigate how three top-performing SOTA detectors behave in new ternary classification context. Our results highlight why "undecided" category is much needed from the viewpoint of explainability. Additionally, we conducted an analysis of explainability of the three best-performing detectors and the explanation notes of the human annotators, revealing insights about the complexity of explainable detection of machine-generated texts. Finally, we propose guidelines for developing future detection systems with improved explanatory power., Comment: 19 pages, 2 figures

Retrieval-augmented generation has gained popularity as a framework to enhance large language models with external knowledge. However, its effectiveness hinges on the retrieval robustness of the model. If the model lacks retrieval robustness, its performance is constrained by the accuracy of the retriever, resulting in significant compromises when the retrieved context is irrelevant. In this paper, we evaluate the "implicit" retrieval robustness of various large language models, instructing them to directly output the final answer without explicitly judging the relevance of the retrieved context. Our findings reveal that fine-tuning on a mix of gold and distracting context significantly enhances the model's robustness to retrieval inaccuracies, while still maintaining its ability to extract correct answers when retrieval is accurate. This suggests that large language models can implicitly handle relevant or irrelevant retrieved context by learning solely from the supervision of the final answer in an end-to-end manner. Introducing an additional process for explicit relevance judgment can be unnecessary and disrupts the end-to-end approach.

Application Programming Interface (API) incompatibility is a long-standing issue in Android application development. The rapid evolution of Android APIs results in a significant number of API additions, removals, and changes between adjacent versions. Unfortunately, this high frequency of alterations may lead to compatibility issues, often without adequate notification to developers regarding these changes. Although researchers have proposed some work on detecting compatibility issues caused by changes in API signatures, they often overlook compatibility issues stemming from sophisticated semantic changes. In response to this challenge, we conducted a large-scale discovery of incompatible APIs in the Android Open Source Project (AOSP) by leveraging static analysis and pre-trained Large Language Models (LLMs) across adjacent versions. We systematically formulate the problem and propose a unified framework to detect incompatible APIs, especially for semantic changes. It's worth highlighting that our approach achieves a 0.83 F1-score in identifying semantically incompatible APIs in the Android framework. Ultimately, our approach detects 5,481 incompatible APIs spanning from version 4 to version 33. We further demonstrate its effectiveness in supplementing the state-of-the-art methods in detecting a broader spectrum of compatibility issues (+92.3%) that have been previously overlooked.

DBSCAN and OPTICS are powerful algorithms for identifying clusters of points in domains where few assumptions can be made about the structure of the data. In this paper, we leverage these strengths and introduce a new algorithm, LINSCAN, designed to seek lineated clusters that are difficult to find and isolate with existing methods. In particular, by embedding points as normal distributions approximating their local neighborhoods and leveraging a distance function derived from the Kullback Leibler Divergence, LINSCAN can detect and distinguish lineated clusters that are spatially close but have orthogonal covariances. We demonstrate how LINSCAN can be applied to seismic data to identify active faults, including intersecting faults, and determine their orientation. Finally, we discuss the properties a generalization of DBSCAN and OPTICS must have in order to retain the stability benefits of these algorithms.

Intertwining of different quantum degrees of freedom manifests exotic quantum phenomena in many-body systems, especially in reduced dimensionality. Here we show that monolayered NbTe2 serves as an ideal platform where lattice, charge, and spin degrees of freedom manifest cooperatively, leading to a new and threading order of chirality. By using spin-polarized scanning tunneling microscopy/spectroscopy, we reveal that the root19 * root19 phase of NbTe2 is encoded with both alternating-chiral atomic displacements and charge density waves, characterized by two chiral units of opposite handedness within the reconstructed cell. We show unambiguous evidence for emergent spin polarizations spreading over the primitive cell, with the magnetization orientation synchronized with alternating handedness of chiral order. Our first-principles studies identify the origin of intertwined orders being correlation driven, with the threading order of chirality emerging when the on-site Coulomb repulsion exceeds a critical value. The spin ordering is further shown to be of hybrid ferrimagnetic nature, contributed by the itinerant electrons and localized d-orbitals. Collectively, these findings expand the realm of chiral order in correlated electron systems, and facilitate an appealing platform for chiral spintronic and related applications.

Diffusion models (DMs) have gained attention in Missing Data Imputation (MDI), but there remain two long-neglected issues to be addressed: (1). Inaccurate Imputation, which arises from inherently sample-diversification-pursuing generative process of DMs. (2). Difficult Training, which stems from intricate design required for the mask matrix in model training stage. To address these concerns within the realm of numerical tabular datasets, we introduce a novel principled approach termed Kernelized Negative Entropy-regularized Wasserstein gradient flow Imputation (KnewImp). Specifically, based on Wasserstein gradient flow (WGF) framework, we first prove that issue (1) stems from the cost functionals implicitly maximized in DM-based MDI are equivalent to the MDI's objective plus diversification-promoting non-negative terms. Based on this, we then design a novel cost functional with diversification-discouraging negative entropy and derive our KnewImp approach within WGF framework and reproducing kernel Hilbert space. After that, we prove that the imputation procedure of KnewImp can be derived from another cost functional related to the joint distribution, eliminating the need for the mask matrix and hence naturally addressing issue (2). Extensive experiments demonstrate that our proposed KnewImp approach significantly outperforms existing state-of-the-art methods.

We present a benchmark for methods in causal learning. Specifically, we consider training a rich class of causal models from time-series data, and we suggest the use of the Krebs cycle and models of metabolism more broadly.

While large language models (LLMs) have showcased impressive capabilities, they struggle with addressing legal queries due to the intricate complexities and specialized expertise required in the legal field. In this paper, we introduce InternLM-Law, a specialized LLM tailored for addressing diverse legal queries related to Chinese laws, spanning from responding to standard legal questions (e.g., legal exercises in textbooks) to analyzing complex real-world legal situations. We meticulously construct a dataset in the Chinese legal domain, encompassing over 1 million queries, and implement a data filtering and processing pipeline to ensure its diversity and quality. Our training approach involves a novel two-stage process: initially fine-tuning LLMs on both legal-specific and general-purpose content to equip the models with broad knowledge, followed by exclusive fine-tuning on high-quality legal data to enhance structured output generation. InternLM-Law achieves the highest average performance on LawBench, outperforming state-of-the-art models, including GPT-4, on 13 out of 20 subtasks. We make InternLM-Law and our dataset publicly available to facilitate future research in applying LLMs within the legal domain., Comment: Our dataset, code and models will be released at https://github.com/InternLM/InternLM-Law

Counter-narratives, which are direct responses consisting of non-aggressive fact-based arguments, have emerged as a highly effective approach to combat the proliferation of hate speech. Previous methodologies have primarily focused on fine-tuning and post-editing techniques to ensure the fluency of generated contents, while overlooking the critical aspects of individualization and relevance concerning the specific hatred targets, such as LGBT groups, immigrants, etc. This research paper introduces a novel framework based on contrastive optimal transport, which effectively addresses the challenges of maintaining target interaction and promoting diversification in generating counter-narratives. Firstly, an Optimal Transport Kernel (OTK) module is leveraged to incorporate hatred target information in the token representations, in which the comparison pairs are extracted between original and transported features. Secondly, a self-contrastive learning module is employed to address the issue of model degeneration. This module achieves this by generating an anisotropic distribution of token representations. Finally, a target-oriented search method is integrated as an improved decoding strategy to explicitly promote domain relevance and diversification in the inference process. This strategy modifies the model's confidence score by considering both token similarity and target relevance. Quantitative and qualitative experiments have been evaluated on two benchmark datasets, which demonstrate that our proposed model significantly outperforms current methods evaluated by metrics from multiple aspects., Comment: IEEE jounrnals

Large language models (LLMs) have significantly advanced various natural language processing tasks, but deploying them remains computationally expensive. Knowledge distillation (KD) is a promising solution, enabling the transfer of capabilities from larger teacher LLMs to more compact student models. Particularly, sequence-level KD, which distills rationale-based reasoning processes instead of merely final outcomes, shows great potential in enhancing students' reasoning capabilities. However, current methods struggle with sequence level KD under long-tailed data distributions, adversely affecting generalization on sparsely represented domains. We introduce the Multi-Stage Balanced Distillation (BalDistill) framework, which iteratively balances training data within a fixed computational budget. By dynamically selecting representative head domain examples and synthesizing tail domain examples, BalDistill achieves state-of-the-art performance across diverse long-tailed datasets, enhancing both the efficiency and efficacy of the distilled models., Comment: preprint

With deep learning (DL) technology becoming an integral part of the new intelligent software, tools of DL framework testing and bug-finding are in high demand. Existing DL framework testing tools have limited coverage on bug types. For example, they lack the capability of finding performance bugs, which are critical for DL model training and inference regarding performance, economics, and the environment. This problem is challenging due to the difficulty of getting test oracles of performance bugs. Moreover, existing tools are inefficient, generating hundreds of test cases with few trigger bugs. In this paper, we propose CITADEL, a method that accelerates the finding of bugs in terms of efficiency and effectiveness. We observe that many DL framework bugs are similar due to the similarity of operators and algorithms belonging to the same family (e.g., Conv2D and Conv3D). Orthogonal to existing bug-finding tools, CITADEL aims to find new bugs that are similar to reported ones that have known test oracles. It works by first collecting existing bug reports and identifying problematic APIs. CITADEL defines context similarity to measure the similarity of DL framework API pairs and automatically generates test cases with oracles for APIs that are similar to the problematic APIs in existing bug reports. CITADEL respectively covers 1,436 PyTorch and 5,380 TensorFlow APIs and effectively detects 79 and 80 API bugs, among which 58 and 68 are new, and 36 and 58 have been confirmed, many of which, e.g., the 11 performance bugs cannot be detected by existing tools. Moreover, a remarkable 35.40% of the test cases generated by CITADEL can trigger bugs, which significantly transcends the ratios of 0.74%, 1.23%, and 3.90% exhibited by the state-of-the-art methods, DocTer, DeepREL, and TitanFuzz., Comment: 12 pages, 10 figures

As a fundamental structural feature, the symmetry of materials determines the exotic quantum properties in transition metal dichalcogenides (TMDs) with charge density wave (CDW). Breaking the inversion symmetry, the Janus structure, an artificially constructed lattice, provides an opportunity to tune the CDW states and the related properties. However, limited by the difficulties in atomic-level fabrication and material stability, the experimental visualization of the CDW states in 2D TMDs with Janus structure is still rare. Here, using surface selenization of VTe2, we fabricated monolayer Janus VTeSe. With scanning tunneling microscopy, an unusual root13-root13 CDW state with threefold rotational symmetry breaking was observed and characterized. Combined with theoretical calculations, we find this CDW state can be attributed to the charge modulation in the Janus VTeSe, beyond the conventional electron-phonon coupling. Our findings provide a promising platform for studying the CDW states and artificially tuning the electronic properties toward the applications.

We study the impact of top SiO2 interlayer thickness on the memory window (MW) of Si channel ferroelectric field-effect transistor (FeFET) with TiN/SiO2/Hf0.5Zr0.5O2/SiOx/Si (MIFIS) gate structure. We find that the MW increases with the increasing thickness of the top SiO2 interlayer, and such an increase exhibits a two-stage linear dependence. The physical origin is the presence of the different interfacial charges trapped at the top SiO2/Hf0.5Zr0.5O2 interface. Moreover, we investigate the dependence of endurance characteristics on initial MW. We find that the endurance characteristic degrades with increasing the initial MW. By inserting a 3.4 nm SiO2 dielectric interlayer between the gate metal TiN and the ferroelectric Hf0.5Zr0.5O2, we achieve a MW of 6.3 V and retention over 10 years. Our work is helpful in the device design of FeFET., Comment: 6 pages, 12 figures. arXiv admin note: substantial text overlap with arXiv:2404.15825

Large vision-language models (LVLMs) hallucinate: certain context cues in an image may trigger the language module's overconfident and incorrect reasoning on abnormal or hypothetical objects. Though a few benchmarks have been developed to investigate LVLM hallucinations, they mainly rely on hand-crafted corner cases whose fail patterns may hardly generalize, and finetuning on them could undermine their validity. These motivate us to develop the first automatic benchmark generation approach, AUTOHALLUSION, that harnesses a few principal strategies to create diverse hallucination examples. It probes the language modules in LVLMs for context cues and uses them to synthesize images by: (1) adding objects abnormal to the context cues; (2) for two co-occurring objects, keeping one and excluding the other; or (3) removing objects closely tied to the context cues. It then generates image-based questions whose ground-truth answers contradict the language module's prior. A model has to overcome contextual biases and distractions to reach correct answers, while incorrect or inconsistent answers indicate hallucinations. AUTOHALLUSION enables us to create new benchmarks at the minimum cost and thus overcomes the fragility of hand-crafted benchmarks. It also reveals common failure patterns and reasons, providing key insights to detect, avoid, or control hallucinations. Comprehensive evaluations of top-tier LVLMs, e.g., GPT-4V(ision), Gemini Pro Vision, Claude 3, and LLaVA-1.5, show a 97.7% and 98.7% success rate of hallucination induction on synthetic and real-world datasets of AUTOHALLUSION, paving the way for a long battle against hallucinations.

The recent observation of the fractional quantum anomalous Hall effect in moir\'e fractional Chern insulators (FCI) provides opportunities for investigating zero magnetic field anyons. So far, both experimental and theoretical results suggest that filling > 1/3 FCI states in the first Chern band share features with those of the lowest Landau level (LL). To create the possibility of realizing non-Abelian anyons, one route is to engineer higher flat Chern bands that mimic higher LLs. Here, we investigate the interaction, topology, and ferromagnetism of the second moir\'e miniband in twisted MoTe2 bilayer (tMoTe2). Around filling factor v = -3, i.e., half-filling of the second miniband, we uncover spontaneous ferromagnetism and an incipient Chern insulator state. By measuring the anomalous Hall effect as a function of twist angle, we find that the Chern numbers (C) of the top two moir\'e flat bands have opposite sign (C = -+1) at twist angles above 3.1{\deg} but the same sign (C = -1) around 2.6{\deg}. This observation is consistent with the recently predicted twist-angle dependent band topology, resulting from the competition between moir\'e ferroelectricity and piezoelectricity. As we increase the magnetic field, only the small twist-angle device (2.6{\deg}) experiences a topological phase transition with an emergent C = -2 state. This is attributed to a Zeeman field-induced band crossing between opposite valleys, with the determined C = -1 for the top two bands. Our results lay a firm foundation for understanding the higher flat Chern bands, which is essential for the prediction or discovery of non-Abelian FCIs., Comment: 24 pages, 4 figures

Since the secrecy rate (SR) performance improvement obtained by secure directional modulation (DM) network is limited, an active intelligent reflective surface (IRS)-assisted DM network is considered to attain a high SR. To address the SR maximization problem, a novel method based on Lagrangian dual transform and closed-form fractional programming algorithm (LDT-CFFP) is proposed, where the solutions to base station (BS) beamforming vectors and IRS reflection coefficient matrix are achieved. However, the computational complexity of LDT-CFFP method is high . To reduce its complexity, a blocked IRS-assisted DM network is designed. To meet the requirements of the network performance, a power allocation (PA) strategy is proposed and adopted in the system. Specifically, the system power between BS and IRS, as well as the transmission power for confidential messages (CM) and artificial noise (AN) from the BS, are allocated separately. Then we put forward null-space projection (NSP) method, maximum-ratio-reflecting (MRR) algorithm and PA strategy (NSP-MRR-PA) to solve the SR maximization problem. The CF solutions to BS beamforming vectors and IRS reflection coefficient matrix are respectively attained via NSP and MRR algorithms. For the PA factors, we take advantage of exhaustive search (ES) algorithm, particle swarm optimization (PSO) and simulated annealing (SA) algorithm to search for the solutions. From simulation results, it is verified that the LDT-CFFP method derives a higher SR gain over NSP-MRR-PA method. For NSP-MRR-PA method, the number of IRS units in each block possesses a significant SR performance. In addition, the application PA strategies, namely ES, PSO, SA methods outperforms the other PA strategies with fixed PA factors., Comment: Directional modulation, active intelligent reflective surface, Lagrangian dual transformation, fractional programming, power allocation

In NGC 2992, a galaxy-scale ionized gas outflow driven by AGN has long been recognized, yet its impact on the host galaxy has remained elusive. In this paper, we utilize data from the archival Very Large Telescope (VLT)/MUSE to present a spatially resolved analysis of stellar populations in this galaxy. Two different stellar population templates are employed to fit the stellar continuum, allowing us to determine the light-weighted stellar age, metallicity, the fraction of the young stellar population (age $<100$ Myr, $P_{\rm Y}$), and the average age and metallicity of $P_{\rm Y}$. Our results reveal the presence of a very young stellar population ($\leq40$ Myr) within the dust lane and nearly along the galaxy's major axis. The light-weighted stellar age and the fraction of $P_{\rm Y}$ show negative trends along the major and minor axes. The average age and metallicity of $P_{\rm Y}$ present positive trends with increasing distance, except along the northern direction of the major axis. Within the circumnuclear region ($<1$ kpc), the distribution of the young stellar population is spatially anti-correlated with the AGN outflow cone. The highest fraction of $P_{\rm Y}$ is observed at the outskirts of the nuclear radio bubble in the northern region near the nucleus. Considering the coupling efficiency and timescales, we propose that the AGN outflow in this galaxy may exert both negative and positive feedback on its host. Additionally, the star formation and the AGN activities could be attributed to the interaction between NGC 2992 and NGC 2993., Comment: 15 pages, 9 figures, accepted for publication in ApJ

Efficient electric manipulation of valley degrees of freedom is critical and challenging for the advancement of valley-based information science and technology. We put forth an electrical scheme, based on a two-band Dirac model, that can switch the fully valley-polarized photocurrent between K and K' valleys using the circular bulk electro-photovoltaic effect. This is accomplished by applying an out-of-plane electric field to the two-dimensional valley materials, which enables continuous tuning of the Berry curvature and its sign flip. We found that the switch of the fully valley-polarized photocurrent is directly tied to the sign change of Berry curvature, which accompanies a topological phase transition, for instance, the quantum spin Hall effect and the quantum valley Hall effect. This scheme has been confirmed in monolayer BiAsI2 and germanene through first-principles calculations. Our paper offers a promising strategy for the development of a volatile valley-addressable memory device and could inspire further research in this area.

Multi-static integrated sensing and communication (ISAC) technology, which can achieve a wider coverage range and avoid self-interference, is an important trend for the future development of ISAC. Existing multi-static ISAC designs are unable to support the asymmetric uplink (UL)/downlink (DL) communication requirements in the scenario while simultaneously achieving optimal sensing performance. This paper proposes a design for multi-static ISAC based on network-assisted full-duplex (NAFD) cell-free networks can well solve the above problems. Under this design, closed-form expressions for the individual comunication rate and localization error rate are derived under imperfect channel state information, which are respectively utilized to assess the communication and sensing performances. Then, we propose a deep Q-network-based accesss point (AP) duplex mode optimization algorithm to obtain the trade-off between communication and sensing from the UL and DL perspectives of the APs. Simulation results demonstrate that the NAFD-based ISAC system proposed in this paper can achieve significantly better communication performance than other ISAC systems while ensuring minimal impact on sensing performance. Then, we validate the accuracy of the derived closed-form expressions. Furthermore, the proposed optimization algorithm achieves performance comparable to that of the exhaustion method with low complexity.

Document translation poses a challenge for Neural Machine Translation (NMT) systems. Most document-level NMT systems rely on meticulously curated sentence-level parallel data, assuming flawless extraction of text from documents along with their precise reading order. These systems also tend to disregard additional visual cues such as the document layout, deeming it irrelevant. However, real-world documents often possess intricate text layouts that defy these assumptions. Extracting information from Optical Character Recognition (OCR) or heuristic rules can result in errors, and the layout (e.g., paragraphs, headers) may convey relationships between distant sections of text. This complexity is particularly evident in widely used PDF documents, which represent information visually. This paper addresses this gap by introducing M3T, a novel benchmark dataset tailored to evaluate NMT systems on the comprehensive task of translating semi-structured documents. This dataset aims to bridge the evaluation gap in document-level NMT systems, acknowledging the challenges posed by rich text layouts in real-world applications., Comment: NAACL 2024, dataset at https://github.com/amazon-science/m3t-multi-modal-translation-bench

An optimization-based image reconstruction algorithm is developed for contrast enhanced digital breast tomosynthesis (DBT) using dual-energy scanning. The algorithm minimizes directional total variation (TV) with a data discrepancy and non-negativity constraints. Iodinated contrast agent (ICA) imaging is performed by reconstructing images from dual-energy DBT data followed by weighted subtraction. Physical DBT data is acquired with a Siemens Mammomat scanner of a structured breast phantom with ICA inserts. Results are shown for both directional TV minimization and filtered back-projection for reference. It is seen that directional TV is able to substantially reduce depth blur for the ICA objects., Comment: Proceedings paper for accepted contribution to the 8th International Conference on Image Formation in X-Ray Computed Tomography (https://www.ct-meeting.org)

Highly crystalline and easily feasible topological insulator-superconductor (TI-SC) heterostructures are crucial for the development of practical topological qubit devices. The optimal superconducting layer for TI-SC heterostructures should be highly resilient against external contaminations and structurally compatible with TIs. In this study, we provide a solution to this challenge by showcasing the growth of a highly crystalline TI-SC heterostructure using refractory TiN (111) as the superconducting layer. This approach can eliminate the need for in-situ cleaving or growth. More importantly, the TiN surface shows high resilience against contaminations during air exposure, as demonstrated by the successful recyclable growth of Bi2Se3. Our findings indicate that TI-SC heterostructures based on nitride films are compatible with device fabrication techniques, paving a path to the realization of practical topological qubit devices in the future., Comment: 22 pages, 4 figures, accepted by ACS Applied Materials & Interfaces

Scene Graph Generation (SGG) is a high-level visual understanding and reasoning task aimed at extracting entities (such as objects) and their interrelationships from images. Significant progress has been made in the study of SGG in natural images in recent years, but its exploration in the domain of remote sensing images remains very limited. The complex characteristics of remote sensing images necessitate higher time and manual interpretation costs for annotation compared to natural images. The lack of a large-scale public SGG benchmark is a major impediment to the advancement of SGG-related research in aerial imagery. In this paper, we introduce the first publicly available large-scale, million-level relation dataset in the field of remote sensing images which is named as ReCon1M. Specifically, our dataset is built upon Fair1M and comprises 21,392 images. It includes annotations for 859,751 object bounding boxes across 60 different categories, and 1,149,342 relation triplets across 64 categories based on these bounding boxes. We provide a detailed description of the dataset's characteristics and statistical information. We conducted two object detection tasks and three sub-tasks within SGG on this dataset, assessing the performance of mainstream methods on these tasks.

The large language models (LLMs), which bridge the gap between human language understanding and complex problem-solving, achieve state-of-the-art performance on several NLP tasks, particularly in few-shot and zero-shot settings. Despite the demonstrable efficacy of LMMs, due to constraints on computational resources, users have to engage with open-source language models or outsource the entire training process to third-party platforms. However, research has demonstrated that language models are susceptible to potential security vulnerabilities, particularly in backdoor attacks. Backdoor attacks are designed to introduce targeted vulnerabilities into language models by poisoning training samples or model weights, allowing attackers to manipulate model responses through malicious triggers. While existing surveys on backdoor attacks provide a comprehensive overview, they lack an in-depth examination of backdoor attacks specifically targeting LLMs. To bridge this gap and grasp the latest trends in the field, this paper presents a novel perspective on backdoor attacks for LLMs by focusing on fine-tuning methods. Specifically, we systematically classify backdoor attacks into three categories: full-parameter fine-tuning, parameter-efficient fine-tuning, and attacks without fine-tuning. Based on insights from a substantial review, we also discuss crucial issues for future research on backdoor attacks, such as further exploring attack algorithms that do not require fine-tuning, or developing more covert attack algorithms.

We present the new concept of photonic alloy as a non-periodic topological material. By mixing non-magnetized and magnetized rods in a non-periodic 2D photonic crystal configuration, we realized photonic alloys in the microwave regime. Our experimental findings reveal that the photonic alloy sustains non-reciprocal chiral edge states (CESs) even at very low concentration of magnetized rods. The non-trivial topology and the associated edge states of these non-periodic systems can be characterized by the winding of the reflection phase. Our results indicate that the threshold concentrations for the investigated system within the first non-trivial band gap to exhibit topological behavior approach zero in the thermodynamic limit for substitutional alloys, while the threshold remains non-zero for interstitial alloys. At low concentration, the system exhibits an inhomogeneous structure characterized by isolated patches of non-percolating magnetic domains that are spaced far apart within a topologically trivial photonic crystal. Surprisingly, the system manifests CESs despite a local breakdown of time-reversal symmetry rather than a global one. Photonic alloys represent a new category of disordered topological materials, offering exciting opportunities for exploring topological materials with adjustable gaps., Comment: 6 pages, 4 figures

The explainability of recommender systems has attracted significant attention in academia and industry. Many efforts have been made for explainable recommendations, yet evaluating the quality of the explanations remains a challenging and unresolved issue. In recent years, leveraging LLMs as evaluators presents a promising avenue in Natural Language Processing tasks (e.g., sentiment classification, information extraction), as they perform strong capabilities in instruction following and common-sense reasoning. However, evaluating recommendation explanatory texts is different from these NLG tasks, as its criteria are related to human perceptions and are usually subjective. In this paper, we investigate whether LLMs can serve as evaluators of recommendation explanations. To answer the question, we utilize real user feedback on explanations given from previous work and additionally collect third-party annotations and LLM evaluations. We design and apply a 3-level meta evaluation strategy to measure the correlation between evaluator labels and the ground truth provided by users. Our experiments reveal that LLMs, such as GPT4, can provide comparable evaluations with appropriate prompts and settings. We also provide further insights into combining human labels with the LLM evaluation process and utilizing ensembles of multiple heterogeneous LLM evaluators to enhance the accuracy and stability of evaluations. Our study verifies that utilizing LLMs as evaluators can be an accurate, reproducible and cost-effective solution for evaluating recommendation explanation texts. Our code is available at https://github.com/Xiaoyu-SZ/LLMasEvaluator.

Speech restoration aims at restoring high quality speech in the presence of a diverse set of distortions. Although several deep learning paradigms have been studied for this task, the power of the recently emerging language models has not been fully explored. In this paper, we propose MaskSR, a masked language model capable of restoring full-band 44.1 kHz speech jointly considering noise, reverb, clipping, and low bandwidth. MaskSR works with discrete acoustic tokens extracted using a pre-trained neural codec. During training, MaskSR is optimized to predict randomly masked tokens extracted from the high quality target speech, conditioned on the corrupted speech with various distortions. During inference, MaskSR reconstructs the target speech tokens with efficient iterative sampling. Extensive experiments show that MaskSR obtains competitive results on both the full-band speech restoration task and also on sub-tasks compared with a wide range of models., Comment: Accepted by INTERSPEECH 2024. Demo page: https://masksr.github.io/MaskSR/

While Large Language Models (LLMs) have achieved remarkable success in various fields, the efficiency of training and inference remains a major challenge. To address this issue, we propose SUBLLM, short for Subsampling-Upsampling-Bypass Large Language Model, an innovative architecture that extends the core decoder-only framework by incorporating subsampling, upsampling, and bypass modules. The subsampling modules are responsible for shortening the sequence, while the upsampling modules restore the sequence length, and the bypass modules enhance convergence. In comparison to LLaMA, the proposed SUBLLM exhibits significant enhancements in both training and inference speeds as well as memory usage, while maintaining competitive few-shot performance. During training, SUBLLM increases speeds by 26% and cuts memory by 10GB per GPU. In inference, it boosts speeds by up to 37% and reduces memory by 1GB per GPU. The training and inference speeds can be enhanced by 34% and 52% respectively when the context window is expanded to 8192. Our code is available at https://github.com/XiaoMi/subllm., Comment: 10 pages, 5 figures, accepted by ECAI 2024

As robotics continues to advance, the need for adaptive and continuously-learning embodied agents increases, particularly in the realm of assistance robotics. Quick adaptability and long-term information retention are essential to operate in dynamic environments typical of humans' everyday lives. A lifelong learning paradigm is thus required, but it is scarcely addressed by current robotics literature. This study empirically investigates the impact of catastrophic forgetting and the effectiveness of knowledge transfer in neural networks trained continuously in an embodied setting. We focus on the task of visual odometry, which holds primary importance for embodied agents in enabling their self-localization. We experiment on the simple continual scenario of discrete transitions between indoor locations, akin to a robot navigating different apartments. In this regime, we observe initial satisfactory performance with high transferability between environments, followed by a specialization phase where the model prioritizes current environment-specific knowledge at the expense of generalization. Conventional regularization strategies and increased model capacity prove ineffective in mitigating this phenomenon. Rehearsal is instead mildly beneficial but with the addition of a substantial memory cost. Incorporating action information, as commonly done in embodied settings, facilitates quicker convergence but exacerbates specialization, making the model overly reliant on its motion expectations and less adept at correctly interpreting visual cues. These findings emphasize the open challenges of balancing adaptation and memory retention in lifelong robotics and contribute valuable insights into the application of a lifelong paradigm on embodied agents., Comment: Accepted to CoLLAs 2024

In this technical report, we introduce the training methodologies implemented in the development of Skywork-MoE, a high-performance mixture-of-experts (MoE) large language model (LLM) with 146 billion parameters and 16 experts. It is initialized from the pre-existing dense checkpoints of our Skywork-13B model. We explore the comparative effectiveness of upcycling versus training from scratch initializations. Our findings suggest that the choice between these two approaches should consider both the performance of the existing dense checkpoints and the MoE training budget. We highlight two innovative techniques: gating logit normalization, which improves expert diversification, and adaptive auxiliary loss coefficients, allowing for layer-specific adjustment of auxiliary loss coefficients. Our experimental results validate the effectiveness of these methods. Leveraging these techniques and insights, we trained our upcycled Skywork-MoE on a condensed subset of our SkyPile corpus. The evaluation results demonstrate that our model delivers strong performance across a wide range of benchmarks.