Search

Large language models (LLMs) have shown impressive capabilities across various natural language tasks. However, evaluating their alignment with human preferences remains a challenge. To this end, we propose a comprehensive human evaluation framework to assess LLMs' proficiency in following instructions on diverse real-world tasks. We construct a hierarchical task tree encompassing 7 major areas covering over 200 categories and over 800 tasks, which covers diverse capabilities such as question answering, reasoning, multiturn dialogue, and text generation, to evaluate LLMs in a comprehensive and in-depth manner. We also design detailed evaluation standards and processes to facilitate consistent, unbiased judgments from human evaluators. A test set of over 3,000 instances is released, spanning different difficulty levels and knowledge domains. Our work provides a standardized methodology to evaluate human alignment in LLMs for both English and Chinese. We also analyze the feasibility of automating parts of evaluation with a strong LLM (GPT-4). Our framework supports a thorough assessment of LLMs as they are integrated into real-world applications. We have made publicly available the task tree, TencentLLMEval dataset, and evaluation methodology which have been demonstrated as effective in assessing the performance of Tencent Hunyuan LLMs. By doing so, we aim to facilitate the benchmarking of advances in the development of safe and human-aligned LLMs.

Monaural speech enhancement has achieved remarkable progress recently. However, its performance has been constrained by the limited spatial cues available at a single microphone. To overcome this limitation, we introduce a strategy to map monaural speech into a fixed simulation space for better differentiation between target speech and noise. Concretely, we propose SE-TerrNet, a novel monaural speech enhancement model featuring a virtual binaural speech mapping network via a two-stage multi-task learning framework. In the first stage, monaural noisy input is projected into a virtual space using supervised speech mapping blocks, creating binaural representations. These blocks synthesize binaural noisy speech from monaural input via an ideal binaural room impulse response. The synthesized output assigns speech and noise sources to fixed directions within the perceptual space. In the second stage, the obtained binaural features from the first stage are aggregated. This aggregation aims to decrease pattern discrepancies between the mapped binaural and original monaural features, achieved by implementing an intermediate fusion module. Furthermore, this stage incorporates the utilization of cross-attention to capture the injected virtual spatial information to improve the extraction of the target speech. Empirical studies highlight the effectiveness of virtual spatial cues in enhancing monaural speech enhancement. As a result, the proposed SE-TerrNet significantly surpasses the recent monaural speech enhancement methods in terms of both speech quality and intelligibility.

Representation learning is increasingly employed to generate representations that are predictive across multiple downstream tasks. The development of representation learning algorithms that provide strong fairness guarantees is thus important because it can prevent unfairness towards disadvantaged groups for all downstream prediction tasks. To prevent unfairness towards disadvantaged groups in all downstream tasks, it is crucial to provide representation learning algorithms that provide fairness guarantees. In this paper, we formally define the problem of learning representations that are fair with high confidence. We then introduce the Fair Representation learning with high-confidence Guarantees (FRG) framework, which provides high-confidence guarantees for limiting unfairness across all downstream models and tasks, with user-defined upper bounds. After proving that FRG ensures fairness for all downstream models and tasks with high probability, we present empirical evaluations that demonstrate FRG's effectiveness at upper bounding unfairness for multiple downstream models and tasks.

Massive multiple-input multiple-output (MIMO) has gained widespread popularity in recent years due to its ability to increase data rates, improve signal quality, and provide better coverage in challenging environments. In this paper, we investigate the MIMO beam selection (MBS) problem, which is proven to be NP-hard and computationally intractable. To deal with this problem, quantum computing that can provide faster and more efficient solutions to large-scale combinatorial optimization is considered. MBS is formulated in a quadratic unbounded binary optimization form and solved with Coherent Ising Machine (CIM) physical machine. We compare the performance of our solution with two classic heuristics, simulated annealing and Tabu search. The results demonstrate an average performance improvement by a factor of 261.23 and 20.6, respectively, which shows that CIM-based solution performs significantly better in terms of selecting the optimal subset of beams. This work shows great promise for practical 5G operation and promotes the application of quantum computing in solving computationally hard problems in communication., Comment: Accepted by IEEE Globecom 2023

The rapid development of intelligent transportation systems and connected vehicles has highlighted the need for secure and efficient key management systems (KMS). In this paper, we introduce VDKMS (Vehicular Decentralized Key Management System), a novel Decentralized Key Management System designed specifically as an infrastructure for Cellular Vehicular-to-Everything (V2X) networks, utilizing a blockchain-based approach. The proposed VDKMS addresses the challenges of secure communication, privacy preservation, and efficient key management in V2X scenarios. It integrates blockchain technology, Self-Sovereign Identity (SSI) principles, and Decentralized Identifiers (DIDs) to enable secure and trustworthy V2X applications among vehicles, infrastructures, and networks. We first provide a comprehensive overview of the system architecture, components, protocols, and workflows, covering aspects such as provisioning, registration, verification, and authorization. We then present a detailed performance evaluation, discussing the security properties and compatibility of the proposed solution, as well as a security analysis. Finally, we present potential applications in the vehicular ecosystem that can leverage the advantages of our approach., Comment: 6 pages, 6 figures, accepted by IEEE Globecom 2023

Goal-conditioned rearrangement of deformable objects (e.g. straightening a rope and folding a cloth) is one of the most common deformable manipulation tasks, where the robot needs to rearrange a deformable object into a prescribed goal configuration with only visual observations. These tasks are typically confronted with two main challenges: the high dimensionality of deformable configuration space and the underlying complexity, nonlinearity and uncertainty inherent in deformable dynamics. To address these challenges, we propose a novel representation strategy that can efficiently model the deformable object states with a set of keypoints and their interactions. We further propose local-graph neural network (GNN), a light local GNN learning to jointly model the deformable rearrangement dynamics and infer the optimal manipulation actions (e.g. pick and place) by constructing and updating two dynamic graphs. Both simulated and real experiments have been conducted to demonstrate that the proposed dynamic graph representation shows superior expressiveness in modeling deformable rearrangement dynamics. Our method reaches much higher success rates on a variety of deformable rearrangement tasks (96.3% on average) than state-of-the-art method in simulation experiments. Besides, our method is much more lighter and has a 60% shorter inference time than state-of-the-art methods. We also demonstrate that our method performs well in the multi-task learning scenario and can be transferred to real-world applications with an average success rate of 95% by solely fine tuning a keypoint detector.

Advance Persistent Threats (APTs), adopted by most delicate attackers, are becoming increasing common and pose great threat to various enterprises and institutions. Data provenance analysis on provenance graphs has emerged as a common approach in APT detection. However, previous works have exhibited several shortcomings: (1) requiring attack-containing data and a priori knowledge of APTs, (2) failing in extracting the rich contextual information buried within provenance graphs and (3) becoming impracticable due to their prohibitive computation overhead and memory consumption. In this paper, we introduce MAGIC, a novel and flexible self-supervised APT detection approach capable of performing multi-granularity detection under different level of supervision. MAGIC leverages masked graph representation learning to model benign system entities and behaviors, performing efficient deep feature extraction and structure abstraction on provenance graphs. By ferreting out anomalous system behaviors via outlier detection methods, MAGIC is able to perform both system entity level and batched log level APT detection. MAGIC is specially designed to handle concept drift with a model adaption mechanism and successfully applies to universal conditions and detection scenarios. We evaluate MAGIC on three widely-used datasets, including both real-world and simulated attacks. Evaluation results indicate that MAGIC achieves promising detection results in all scenarios and shows enormous advantage over state-of-the-art APT detection approaches in performance overhead., Comment: 18 pages, 13 figures, to appear in the 33rd USENIX Security Symposium (USENIX Security 2024)

With the recent proliferation of large language models (LLMs), such as Generative Pre-trained Transformers (GPT), there has been a significant shift in exploring human and machine comprehension of semantic language meaning. This shift calls for interdisciplinary research that bridges cognitive science and natural language processing (NLP). This pilot study aims to provide insights into individuals' neural states during a semantic relation reading-comprehension task. We propose jointly analyzing LLMs, eye-gaze, and electroencephalographic (EEG) data to study how the brain processes words with varying degrees of relevance to a keyword during reading. We also use a feature engineering approach to improve the fixation-related EEG data classification while participants read words with high versus low relevance to the keyword. The best validation accuracy in this word-level classification is over 60\% across 12 subjects. Words of high relevance to the inference keyword had significantly more eye fixations per word: 1.0584 compared to 0.6576 when excluding no-fixation words, and 1.5126 compared to 1.4026 when including them. This study represents the first attempt to classify brain states at a word level using LLM knowledge. It provides valuable insights into human cognitive abilities and the realm of Artificial General Intelligence (AGI), and offers guidance for developing potential reading-assisted technologies.

In the past decades, model averaging (MA) has attracted much attention as it has emerged as an alternative tool to the model selection (MS) statistical approach. Hansen [Econometrica 75 (2007) 1175--1189] introduced a Mallows model averaging (MMA) method with model weights selected by minimizing a Mallows' $C_p$ criterion. The main theoretical justification for MMA is an asymptotic optimality (AOP), which states that the risk/loss of the resulting MA estimator is asymptotically equivalent to that of the best but infeasible averaged model. MMA's AOP is proved in the literature by either constraining weights in a special discrete weight set or limiting the number of candidate models. In this work, it is first shown that under these restrictions, however, the optimal risk of MA becomes an unreachable target, and MMA may converge more slowly than MS. In this background, a foundational issue that has not been addressed is: When a suitably large set of candidate models is considered, and the model weights are not harmfully constrained, can the MMA estimator perform asymptotically as well as the optimal convex combination of the candidate models? We answer this question in both nested and non-nested settings. In the nested setting, we provide finite sample inequalities for the risk of MMA and show that without unnatural restrictions on the candidate models, MMA's AOP holds in a general continuous weight set under certain mild conditions. In the non-nested setting, a sufficient condition and a negative result are established for the achievability of the optimal MA risk. Implications on minimax adaptivity are given as well. The results from simulations back up our theoretical findings.

This study explores how sentence types affect the Lombard effect and intelligibility enhancement, focusing on comparisons between natural and grid sentences. Using the Lombard Chinese-TIMIT (LCT) corpus and the Enhanced MAndarin Lombard Grid (EMALG) corpus, we analyze changes in phonetic and acoustic features across different noise levels. Our results show that grid sentences produce more pronounced Lombard effects than natural sentences. Then, we develop and test a normal-to-Lombard conversion model, trained separately on LCT and EMALG corpora. Through subjective and objective evaluations, natural sentences are superior in maintaining speech quality in intelligibility enhancement. In contrast, grid sentences could provide superior intelligibility due to the more pronounced Lombard effect. This study provides a valuable perspective on enhancing speech communication in noisy environments.

Graph Neural Networks (GNNs) have been shown to possess strong representation abilities over graph data. However, GNNs are vulnerable to adversarial attacks, and even minor perturbations to the graph structure can significantly degrade their performance. Existing methods either are ineffective against sophisticated attacks or require the optimization of dense adjacency matrices, which is time-consuming and prone to local minima. To remedy this problem, we propose an Efficient Low-Rank Graph Neural Network (ELR-GNN) defense method, which aims to learn low-rank and sparse graph structures for defending against adversarial attacks, ensuring effective defense with greater efficiency. Specifically, ELR-GNN consists of two modules: a Coarse Low-Rank Estimation Module and a Fine-Grained Estimation Module. The first module adopts the truncated Singular Value Decomposition (SVD) to initialize the low-rank adjacency matrix estimation, which serves as a starting point for optimizing the low-rank matrix. In the second module, the initial estimate is refined by jointly learning a low-rank sparse graph structure with the GNN model. Sparsity is incorporated into the learned low-rank adjacency matrix by pruning weak connections, which can reduce redundant data while maintaining valuable information. As a result, instead of using the dense adjacency matrix directly, ELR-GNN can learn a low-rank and sparse estimate of it in a simple, efficient and easy to optimize manner. The experimental results demonstrate that ELR-GNN outperforms the state-of-the-art GNN defense methods in the literature, in addition to being very efficient and easy to train., Comment: ICKG 2023

Graph Neural Networks (GNNs) require a large number of labeled graph samples to obtain good performance on the graph classification task. The performance of GNNs degrades significantly as the number of labeled graph samples decreases. To reduce the annotation cost, it is therefore important to develop graph augmentation methods that can generate new graph instances to increase the size and diversity of the limited set of available labeled graph samples. In this work, we propose a novel mixup-based graph augmentation method, Graph Dual Mixup (GDM), that leverages both functional and structural information of the graph instances to generate new labeled graph samples. GDM employs a graph structural auto-encoder to learn structural embeddings of the graph samples, and then applies mixup to the structural information of the graphs in the learned structural embedding space and generates new graph structures from the mixup structural embeddings. As for the functional information, GDM applies mixup directly to the input node features of the graph samples to generate functional node feature information for new mixup graph instances. Jointly, the generated input node features and graph structures yield new graph samples which can supplement the set of original labeled graphs. Furthermore, we propose two novel Balanced Graph Sampling methods to enhance the balanced difficulty and diversity for the generated graph samples. Experimental results on the benchmark datasets demonstrate that our proposed method substantially outperforms the state-of-the-art graph augmentation methods when the labeled graphs are scarce., Comment: ECML 2023

The Lombard effect refers to individuals' unconscious modulation of vocal effort in response to variations in the ambient noise levels, intending to enhance speech intelligibility. The impact of different decibel levels and types of background noise on Lombard effects remains unclear. Building upon the characteristic of Lombard speech that individuals adjust their speech to improve intelligibility dynamically based on the self-feedback speech, we propose a flavor classification approach for the Lombard effect. We first collected Mandarin Lombard speech under different noise conditions, then simulated self-feedback speech, and ultimately conducted the statistical test on the word correct rate. We found that both SSN and babble noise types result in four distinct categories of Mandarin Lombard speech in the range of 30 to 80 dBA with different transition points.

This study investigates the Lombard effect, where individuals adapt their speech in noisy environments. We introduce an enhanced Mandarin Lombard grid (EMALG) corpus with meaningful sentences , enhancing the Mandarin Lombard grid (MALG) corpus. EMALG features 34 speakers and improves recording setups, addressing challenges faced by MALG with nonsense sentences. Our findings reveal that in Mandarin, meaningful sentences are more effective in enhancing the Lombard effect. Additionally, we uncover that female exhibit a more pronounced Lombard effect than male when uttering meaningful sentences. Moreover, our results reaffirm the consistency in the Lombard effect comparison between English and Mandarin found in previous research.

3D object detection using LiDAR point clouds is a fundamental task in the fields of computer vision, robotics, and autonomous driving. However, existing 3D detectors heavily rely on annotated datasets, which are both time-consuming and prone to errors during the process of labeling 3D bounding boxes. In this paper, we propose a Scene Completion Pre-training (SCP) method to enhance the performance of 3D object detectors with less labeled data. SCP offers three key advantages: (1) Improved initialization of the point cloud model. By completing the scene point clouds, SCP effectively captures the spatial and semantic relationships among objects within urban environments. (2) Elimination of the need for additional datasets. SCP serves as a valuable auxiliary network that does not impose any additional efforts or data requirements on the 3D detectors. (3) Reduction of the amount of labeled data for detection. With the help of SCP, the existing state-of-the-art 3D detectors can achieve comparable performance while only relying on 20% labeled data., Comment: Wins the best paper award at ISPRS Geospatial Week 2023

A trigger system has been designed and implemented for the HIRFL-CSR external target experiment (CEE), the spectrometer for studying nuclear matter properties with heavy ion collisions in the GeV energy region. The system adopts master-slave structure and serial data transmission mode using optical fiber to deal with different types of detectors and long-distance signal transmission. The trigger logic can be accessed based on command register and controlled by a remote computer. The overall field programmable gate array (FPGA) logic can be flexibly reconfigured online to match the physical requirements of the experiment. The trigger system has been tested in beam experiment. It is demonstrated that the trigger system functions correctly and meets the physical requirements of CEE.

Since rain streaks show a variety of shapes and directions, learning the degradation representation is extremely challenging for single image deraining. Existing methods are mainly targeted at designing complicated modules to implicitly learn latent degradation representation from coupled rainy images. This way, it is hard to decouple the content-independent degradation representation due to the lack of explicit constraint, resulting in over- or under-enhancement problems. To tackle this issue, we propose a novel Latent Degradation Representation Constraint Network (LDRCNet) that consists of Direction-Aware Encoder (DAEncoder), UNet Deraining Network, and Multi-Scale Interaction Block (MSIBlock). Specifically, the DAEncoder is proposed to adaptively extract latent degradation representation by using the deformable convolutions to exploit the direction consistency of rain streaks. Next, a constraint loss is introduced to explicitly constraint the degradation representation learning during training. Last, we propose an MSIBlock to fuse with the learned degradation representation and decoder features of the deraining network for adaptive information interaction, which enables the deraining network to remove various complicated rainy patterns and reconstruct image details. Experimental results on synthetic and real datasets demonstrate that our method achieves new state-of-the-art performance., Comment: This paper is accepted to ICASSP 2024

Thermal metamaterial represents a groundbreaking approach to control heat conduction, and, as a crucial component, thermal invisibility is of utmost importance for heat management. Despite the flourishing development of thermal invisibility schemes, they still face two limitations in practical applications. First, objects are typically completely enclosed in traditional cloaks, making them difficult to use and unsuitable for objects with heat sources. Second, although some theoretical proposals have been put forth to change the thermal conductivity of materials to achieve dynamic invisibility, their designs are complex and rigid, making them unsuitable for large-scale use in real three-dimensional spaces. Here, we propose a concept of a thermal dome to achieve three-dimensional invisibility. Our scheme includes an open functional area, greatly enhancing its usability and applicability. It features a reconfigurable structure, constructed with simple isotropic natural materials, making it suitable for dynamic requirements. The performance of our reconfigurable thermal dome has been confirmed through simulations and experiments, consistent with the theory. The introduction of this concept can greatly advance the development of thermal invisibility technology from theory to engineering and provide inspiration for other physical domains, such as direct current electric fields and magnetic fields.

To investigate the 1.6 GeV high-energy proton beam detector utilized in the CSNS Phase-II upgrade project, a plastic scintillator detector presents a viable option due to its superior radiation hardness. This study investigates the effects of irradiation damage on a BC408 plastic scintillator induced by 80 MeV protons, including absorption and fluorescence spectroscopy, and light yield tests of BC408 pre- and post-proton irradiation, with a focus on determining the radiation resistance threshold of BC408. The results indicate that the performance of BC408 remains unimpaired at absorbed doses up to 5.14*10^3 Gy/cm3, demonstrating its ability to absorb 1.63*10^13 p/cm3 1.6 GeV protons while maintaining stability. This suggests that BC408 could potentially be used as the 1.6 GeV high-energy proton beam detector in the CSNS Phase-II upgrade project.

Cross-domain few-shot classification induces a much more challenging problem than its in-domain counterpart due to the existence of domain shifts between the training and test tasks. In this paper, we develop a novel Adaptive Parametric Prototype Learning (APPL) method under the meta-learning convention for cross-domain few-shot classification. Different from existing prototypical few-shot methods that use the averages of support instances to calculate the class prototypes, we propose to learn class prototypes from the concatenated features of the support set in a parametric fashion and meta-learn the model by enforcing prototype-based regularization on the query set. In addition, we fine-tune the model in the target domain in a transductive manner using a weighted-moving-average self-training approach on the query instances. We conduct experiments on multiple cross-domain few-shot benchmark datasets. The empirical results demonstrate that APPL yields superior performance than many state-of-the-art cross-domain few-shot learning methods.