Search

Weak supervision combines the advantages of training on real data with the ability to exploit signal properties. However, training a neural network using weak supervision often requires an excessive amount of signal data, which severely limits its practical applicability. In this study, we propose addressing this limitation through data augmentation, increasing the training data's size and diversity. Specifically, we focus on physics-inspired data augmentation methods, such as $p_{\text{T}}$ smearing and jet rotation. Our results demonstrate that data augmentation can significantly enhance the performance of weak supervision, enabling neural networks to learn efficiently from substantially less data.

Iterative retrieval refers to the process in which the model continuously queries the retriever during generation to enhance the relevance of the retrieved knowledge, thereby improving the performance of Retrieval-Augmented Generation (RAG). Existing work typically employs few-shot prompting or manually constructed rules to implement iterative retrieval. This introduces additional inference overhead and overlooks the remarkable reasoning capabilities of Large Language Models (LLMs). In this paper, we introduce Auto-RAG, an autonomous iterative retrieval model centered on the LLM's powerful decision-making capabilities. Auto-RAG engages in multi-turn dialogues with the retriever, systematically planning retrievals and refining queries to acquire valuable knowledge. This process continues until sufficient external information is gathered, at which point the results are presented to the user. To this end, we develop a method for autonomously synthesizing reasoning-based decision-making instructions in iterative retrieval and fine-tuned the latest open-source LLMs. The experimental results indicate that Auto-RAG is capable of autonomous iterative interaction with the retriever, effectively leveraging the remarkable reasoning and decision-making abilities of LLMs, which lead to outstanding performance across six benchmarks. Further analysis reveals that Auto-RAG can autonomously adjust the number of iterations based on the difficulty of the questions and the utility of the retrieved knowledge, without requiring any human intervention. Moreover, Auto-RAG expresses the iterative retrieval process in natural language, enhancing interpretability while providing users with a more intuitive experience\footnote{Code is available at \url{https://github.com/ictnlp/Auto-RAG}., Comment: Code is available at https://github.com/ictnlp/Auto-RAG

Streaming generation models are increasingly utilized across various fields, with the Transducer architecture being particularly popular in industrial applications. However, its input-synchronous decoding mechanism presents challenges in tasks requiring non-monotonic alignments, such as simultaneous translation, leading to suboptimal performance in these contexts. In this research, we address this issue by tightly integrating Transducer's decoding with the history of input stream via a learnable monotonic attention mechanism. Our approach leverages the forward-backward algorithm to infer the posterior probability of alignments between the predictor states and input timestamps, which is then used to estimate the context representations of monotonic attention in training. This allows Transducer models to adaptively adjust the scope of attention based on their predictions, avoiding the need to enumerate the exponentially large alignment space. Extensive experiments demonstrate that our MonoAttn-Transducer significantly enhances the handling of non-monotonic alignments in streaming generation, offering a robust solution for Transducer-based frameworks to tackle more complex streaming generation tasks., Comment: Codes: https://github.com/ictnlp/MonoAttn-Transducer

Inference acceleration of large language models (LLMs) has been put forward in many application scenarios and speculative decoding has shown its advantage in addressing inference acceleration. Speculative decoding usually introduces a draft model to assist the base LLM where the draft model produces drafts and the base LLM verifies the draft for acceptance or rejection. In this framework, the final inference speed is decided by the decoding speed of the draft model and the acceptance rate of the draft provided by the draft model. Currently the widely used draft models usually generate draft tokens for the next several positions in a non-autoregressive way without considering the correlations between draft tokens. Therefore, it has a high decoding speed but an unsatisfactory acceptance rate. In this paper, we focus on how to improve the performance of the draft model and aim to accelerate inference via a high acceptance rate. To this end, we propose a CTC-based draft model which strengthens the correlations between draft tokens during the draft phase, thereby generating higher-quality draft candidate sequences. Experiment results show that compared to strong baselines, the proposed method can achieve a higher acceptance rate and hence a faster inference speed.

Large language models (LLMs), with their powerful generative capabilities and vast knowledge, empower various tasks in everyday life. However, these abilities are primarily concentrated in high-resource languages, leaving low-resource languages with weaker generative capabilities and relatively limited knowledge. Enhancing the multilingual capabilities of LLMs is therefore crucial for serving over 100 linguistic communities worldwide. An intuitive approach to enhance the multilingual capabilities would be to construct instruction data for various languages, but constructing instruction data for over 100 languages is prohibitively costly. In this paper, we introduce BayLing 2, which efficiently transfers generative capabilities and knowledge from high-resource languages to low-resource languages through language alignment. To achieve this, we constructed a dataset of 3.2 million instructions, comprising high-resource language instructions (Chinese and English) and cross-lingual instructions for 100+ languages and performed instruction tuning based on the dataset to facilitate the capability transfer between languages. Using Llama as the foundation model, we developed BayLing-2-7B, BayLing-2-13B, and BayLing-2-8B, and conducted a comprehensive evaluation of BayLing. For multilingual translation across 100+ languages, BayLing shows superior performance compared to open-source models of similar scale. For multilingual knowledge and understanding benchmarks, BayLing achieves significant improvements across over 20 low-resource languages, demonstrating its capability of effective knowledge transfer from high-resource to low-resource languages. Furthermore, results on English benchmarks indicate that BayLing maintains high performance in highresource languages while enhancing the performance in low-resource languages. Demo, homepage, code and models of BayLing are available., Comment: BayLing 2's online demo: http://nlp.ict.ac.cn/bayling/demo. BayLing 2's code and models: https://github.com/ictnlp/BayLing

Reconstructing the 3D anatomical structures of the oral cavity, which originally reside in the cone-beam CT (CBCT), from a single 2D Panoramic X-ray(PX) remains a critical yet challenging task, as it can effectively reduce radiation risks and treatment costs during the diagnostic in digital dentistry. However, current methods are either error-prone or only trained/evaluated on small-scale datasets (less than 50 cases), resulting in compromised trustworthiness. In this paper, we propose PX2Tooth, a novel approach to reconstruct 3D teeth using a single PX image with a two-stage framework. First, we design the PXSegNet to segment the permanent teeth from the PX images, providing clear positional, morphological, and categorical information for each tooth. Subsequently, we design a novel tooth generation network (TGNet) that learns to transform random point clouds into 3D teeth. TGNet integrates the segmented patch information and introduces a Prior Fusion Module (PFM) to enhance the generation quality, especially in the root apex region. Moreover, we construct a dataset comprising 499 pairs of CBCT and Panoramic X-rays. Extensive experiments demonstrate that PX2Tooth can achieve an Intersection over Union (IoU) of 0.793, significantly surpassing previous methods, underscoring the great potential of artificial intelligence in digital dentistry., Comment: Ma W, Wu H, Xiao Z, et al. PX2Tooth: Reconstructing the 3D Point Cloud Teeth from a Single Panoramic X-Ray[C]//International Conference on Medical Image Computing and Computer-Assisted Intervention. Cham: Springer Nature Switzerland, 2024: 411-421

Byte-based machine translation systems have shown significant potential in massively multilingual settings. Unicode encoding, which maps each character to specific byte(s), eliminates the emergence of unknown words, even in new languages, enabling broad language scalability. However, byte-level tokenization results in sequences that are hard to interpret due to limited semantic information per byte. Local contextualization has proven effective in assigning initial semantics to tokens, improving sentence comprehension. Nevertheless, variations in encoding rules across languages necessitate an adaptive approach for effective contextualization. To this end, we propose Adaptive MultiScale-Headed Attention (Ada-MSHA), adaptively selecting and mixing attention heads, which are treated as contextualization experts. This enhances the flexibility of contextualization scales and improves the potential to discover a better strategy than previous methods. Experiment results show that our method outperforms existing methods without extensive manual adjustment of hyper-parameters and surpasses subword-based models with fewer parameters in Ted-59 dataset. Our code is available at https://github.com/ictnlp/MoCE.

A dual-operation mode SNSPD is demonstrated. In the conventional Geiger SNSPD mode the sensor operates at temperatures well below the critical temperature, Tc, working as an event counter without sensitivity to the number of photons impinging the sensor. In the calorimetric mode, the detector is operated at temperatures just below Tc and displays photon-number sensitivity for wavelengths in the optical spectrum. In this energy sensitive mode, photon absorption causes Joule heating of the SNSPD that becomes partially resistive without the presence of latching. Depending on the application, by tuning the sample temperature and bias current using the same readout system, the SNSPD can readily switch between the two modes. In the calorimetric mode, SNSPD recovery times shorter than the ones in the Geiger mode are observed, reaching values as low as 580ps. Dual-mode SNSPD's may provide significant advancements in spectroscopy and calorimetry, where precise timing, photon counting and energy resolution are required., Comment: Manuscript prepared for APL

Models like GPT-4o enable real-time interaction with large language models (LLMs) through speech, significantly enhancing user experience compared to traditional text-based interaction. However, there is still a lack of exploration on how to build speech interaction models based on open-source LLMs. To address this, we propose LLaMA-Omni, a novel model architecture designed for low-latency and high-quality speech interaction with LLMs. LLaMA-Omni integrates a pretrained speech encoder, a speech adaptor, an LLM, and a streaming speech decoder. It eliminates the need for speech transcription, and can simultaneously generate text and speech responses directly from speech instructions with extremely low latency. We build our model based on the latest Llama-3.1-8B-Instruct model. To align the model with speech interaction scenarios, we construct a dataset named InstructS2S-200K, which includes 200K speech instructions and corresponding speech responses. Experimental results show that compared to previous speech-language models, LLaMA-Omni provides better responses in both content and style, with a response latency as low as 226ms. Additionally, training LLaMA-Omni takes less than 3 days on just 4 GPUs, paving the way for the efficient development of speech-language models in the future., Comment: Preprint. Project: https://github.com/ictnlp/LLaMA-Omni

The Rust programming language has garnered significant attention due to its robust safety features and memory management capabilities. Despite its guaranteed memory safety, Rust programs suffer from runtime errors that are unmanageable, i.e., panic errors. Notably, traditional memory issues such as null pointer dereferences, which are prevalent in other languages, are less likely to be triggered in Rust due to its strict ownership rules. However, the unique nature of Rust's panic bugs, which arise from the language's stringent safety and ownership paradigms, presents a distinct challenge. Over half of the bugs in rustc, Rust's own compiler, are attributable to crash stemming from panic errors. However, addressing Rust panic bugs is challenging and requires significant effort, as existing fix patterns are not directly applicable due to the design and feature of Rust language.Therefore, developing foundational infrastructure, including datasets, fixing patterns, and automated repair tools, is both critical and urgent. This paper introduces a comprehensive infrastructure, namely PanicFI, aimed at providing supports for understanding Rust panic bugs and developing automated techniques. In PanicFI, we construct a dataset, Panic4R, comprising 102 real panic bugs and their fixes from the top 500 most-downloaded open-source crates.Then, through an analysis of the Rust compiler implementation , we identify Rust-specific patterns for fixing panic bugs, providing insights and guidance for generating patches. Moreover, we develop PanicKiller, the first automated tool for fixing Rust panic bugs, which has already contributed to the resolution of 28 panic bugs in open-source projects.The practicality and efficiency of PanicKiller confirm the effectiveness of the patterns mined within PanicFI.

The many-to-many multilingual neural machine translation can be regarded as the process of integrating semantic features from the source sentences and linguistic features from the target sentences. To enhance zero-shot translation, models need to share knowledge across languages, which can be achieved through auxiliary tasks for learning a universal representation or cross-lingual mapping. To this end, we propose to exploit both semantic and linguistic features between multiple languages to enhance multilingual translation. On the encoder side, we introduce a disentangling learning task that aligns encoder representations by disentangling semantic and linguistic features, thus facilitating knowledge transfer while preserving complete information. On the decoder side, we leverage a linguistic encoder to integrate low-level linguistic features to assist in the target language generation. Experimental results on multilingual datasets demonstrate significant improvement in zero-shot translation compared to the baseline system, while maintaining performance in supervised translation. Further analysis validates the effectiveness of our method in leveraging both semantic and linguistic features. The code is available at https://github.com/ictnlp/SemLing-MNMT., Comment: Accepted by ACL2024 Findings

We propose a functional stochastic block model whose vertices involve functional data information. This new model extends the classic stochastic block model with vector-valued nodal information, and finds applications in real-world networks whose nodal information could be functional curves. Examples include international trade data in which a network vertex (country) is associated with the annual or quarterly GDP over certain time period, and MyFitnessPal data in which a network vertex (MyFitnessPal user) is associated with daily calorie information measured over certain time period. Two statistical tasks will be jointly executed. First, we will detect community structures of the network vertices assisted by the functional nodal information. Second, we propose computationally efficient variational test to examine the significance of the functional nodal information. We show that the community detection algorithms achieve weak and strong consistency, and the variational test is asymptotically chi-square with diverging degrees of freedom. As a byproduct, we propose pointwise confidence intervals for the slop function of the functional nodal information. Our methods are examined through both simulated and real datasets.

Direct speech-to-speech translation (S2ST) has achieved impressive translation quality, but it often faces the challenge of slow decoding due to the considerable length of speech sequences. Recently, some research has turned to non-autoregressive (NAR) models to expedite decoding, yet the translation quality typically lags behind autoregressive (AR) models significantly. In this paper, we investigate the performance of CTC-based NAR models in S2ST, as these models have shown impressive results in machine translation. Experimental results demonstrate that by combining pretraining, knowledge distillation, and advanced NAR training techniques such as glancing training and non-monotonic latent alignments, CTC-based NAR models achieve translation quality comparable to the AR model, while preserving up to 26.81$\times$ decoding speedup., Comment: ACL 2024 Findings

Recently proposed two-pass direct speech-to-speech translation (S2ST) models decompose the task into speech-to-text translation (S2TT) and text-to-speech (TTS) within an end-to-end model, yielding promising results. However, the training of these models still relies on parallel speech data, which is extremely challenging to collect. In contrast, S2TT and TTS have accumulated a large amount of data and pretrained models, which have not been fully utilized in the development of S2ST models. Inspired by this, in this paper, we first introduce a composite S2ST model named ComSpeech, which can seamlessly integrate any pretrained S2TT and TTS models into a direct S2ST model. Furthermore, to eliminate the reliance on parallel speech data, we propose a novel training method ComSpeech-ZS that solely utilizes S2TT and TTS data. It aligns representations in the latent space through contrastive learning, enabling the speech synthesis capability learned from the TTS data to generalize to S2ST in a zero-shot manner. Experimental results on the CVSS dataset show that when the parallel speech data is available, ComSpeech surpasses previous two-pass models like UnitY and Translatotron 2 in both translation quality and decoding speed. When there is no parallel speech data, ComSpeech-ZS lags behind \name by only 0.7 ASR-BLEU and outperforms the cascaded models., Comment: ACL 2024 main conference. Project Page: https://ictnlp.github.io/ComSpeech-Site/

Simultaneous translation models play a crucial role in facilitating communication. However, existing research primarily focuses on text-to-text or speech-to-text models, necessitating additional cascade components to achieve speech-to-speech translation. These pipeline methods suffer from error propagation and accumulate delays in each cascade component, resulting in reduced synchronization between the speaker and listener. To overcome these challenges, we propose a novel non-autoregressive generation framework for simultaneous speech translation (NAST-S2X), which integrates speech-to-text and speech-to-speech tasks into a unified end-to-end framework. We develop a non-autoregressive decoder capable of concurrently generating multiple text or acoustic unit tokens upon receiving fixed-length speech chunks. The decoder can generate blank or repeated tokens and employ CTC decoding to dynamically adjust its latency. Experimental results show that NAST-S2X outperforms state-of-the-art models in both speech-to-text and speech-to-speech tasks. It achieves high-quality simultaneous interpretation within a delay of less than 3 seconds and provides a 28 times decoding speedup in offline generation., Comment: ACL 2024; Codes and demos are at https://github.com/ictnlp/NAST-S2x

Simultaneous Machine Translation (SiMT) generates target translations while reading the source sentence. It relies on a policy to determine the optimal timing for reading sentences and generating translations. Existing SiMT methods generally adopt the traditional Transformer architecture, which concurrently determines the policy and generates translations. While they excel at determining policies, their translation performance is suboptimal. Conversely, Large Language Models (LLMs), trained on extensive corpora, possess superior generation capabilities, but it is difficult for them to acquire translation policy through the training methods of SiMT. Therefore, we introduce Agent-SiMT, a framework combining the strengths of LLMs and traditional SiMT methods. Agent-SiMT contains the policy-decision agent and the translation agent. The policy-decision agent is managed by a SiMT model, which determines the translation policy using partial source sentence and translation. The translation agent, leveraging an LLM, generates translation based on the partial source sentence. The two agents collaborate to accomplish SiMT. Experiments demonstrate that Agent-SiMT attains state-of-the-art performance., Comment: 18 pages, 8 figures, 7 tables. v2 of arXiv:2402.13036

The intrinsic magnetic order, large topological-magnetic gap and rich topological phases make MnBi2Te4 a wonderful platform to study exotic topological quantum states such as axion insulator and Chern insulator. To realize and manipulate these topological phases in a MnBi2Te4 thin film, precise manipulation of the electric field across the film is essential, which requires a dual-gate structure. In this work, we achieve dual-gate tuning of MnBi2Te4 thin films grown with molecular beam epitaxy on SrTiO3(111) substrates by applying the substrate and an AlOx layer as the gate dielectrics of bottom and top gates, respectively. Under magnetic field of 9T and temperature of 20 mK, the Hall and longitudinal resistivities of the films show inversed gate-voltage dependence, for both top- and bottom-gates, signifying the existence of the dissipationless edge state contributed by Chern insulator phase in the ferromagnetic configuration. The maximum of the Hall resistivity only reaches 0.8 h/e2, even with dual-gate tuning, probably due to the high density of bulk carriers introduced by secondary phases. In the antiferromagnetic state under zero magnetic field, the films show normal insulator behavior. The dual-gated MnBi2Te4 thin films lay the foundation for developing devices based on electrically tunable topological quantum states., Comment: 24 pages, 4 figures

Simultaneous Machine Translation (SiMT) generates translation while reading source tokens, essentially producing the target prefix based on the source prefix. To achieve good performance, it leverages the relationship between source and target prefixes to exact a policy to guide the generation of translations. Although existing SiMT methods primarily focus on the Encoder-Decoder architecture, we explore the potential of Decoder-only architecture, owing to its superior performance in various tasks and its inherent compatibility with SiMT. However, directly applying the Decoder-only architecture to SiMT poses challenges in terms of training and inference. To alleviate the above problems, we propose the first Decoder-only SiMT model, named Decoder-only Streaming Transformer (DST). Specifically, DST separately encodes the positions of the source and target prefixes, ensuring that the position of the target prefix remains unaffected by the expansion of the source prefix. Furthermore, we propose a Streaming Self-Attention (SSA) mechanism tailored for the Decoder-only architecture. It is capable of obtaining translation policy by assessing the sufficiency of input source information and integrating with the soft-attention mechanism to generate translations. Experiments demonstrate that our approach achieves state-of-the-art performance on three translation tasks., Comment: Accepted to ACL 2024. 14 pages, 10 Tables, 5 Figures

Simultaneous speech-to-speech translation (Simul-S2ST, a.k.a streaming speech translation) outputs target speech while receiving streaming speech inputs, which is critical for real-time communication. Beyond accomplishing translation between speech, Simul-S2ST requires a policy to control the model to generate corresponding target speech at the opportune moment within speech inputs, thereby posing a double challenge of translation and policy. In this paper, we propose StreamSpeech, a direct Simul-S2ST model that jointly learns translation and simultaneous policy in a unified framework of multi-task learning. Adhering to a multi-task learning approach, StreamSpeech can perform offline and simultaneous speech recognition, speech translation and speech synthesis via an "All-in-One" seamless model. Experiments on CVSS benchmark demonstrate that StreamSpeech achieves state-of-the-art performance in both offline S2ST and Simul-S2ST tasks. Besides, StreamSpeech is able to present high-quality intermediate results (i.e., ASR or translation results) during simultaneous translation process, offering a more comprehensive real-time communication experience., Comment: Accepted to ACL 2024 main conference, Project Page: https://ictnlp.github.io/StreamSpeech-site/

Subword tokenization is a common method for vocabulary building in Neural Machine Translation (NMT) models. However, increasingly complex tasks have revealed its disadvantages. First, a vocabulary cannot be modified once it is learned, making it hard to adapt to new words. Second, in multilingual translation, the imbalance in data volumes across different languages spreads to the vocabulary, exacerbating translations involving low-resource languages. While byte-based tokenization addresses these issues, byte-based models struggle with the low information density inherent in UTF-8 byte sequences. Previous works enhance token semantics through local contextualization but fail to select an appropriate contextualizing scope based on the input. Consequently, we propose the Multi-Scale Contextualization (MSC) method, which learns contextualized information of varying scales across different hidden state dimensions. It then leverages the attention module to dynamically integrate the multi-scale contextualized information. Experiments show that MSC significantly outperforms subword-based and other byte-based methods in both multilingual and out-of-domain scenarios. Code can be found in https://github.com/ictnlp/Multiscale-Contextualization., Comment: Accepted by ACL2024 Findings, renew author's email

Concerns regarding Large Language Models (LLMs) to memorize and disclose private information, particularly Personally Identifiable Information (PII), become prominent within the community. Many efforts have been made to mitigate the privacy risks. However, the mechanism through which LLMs memorize PII remains poorly understood. To bridge this gap, we introduce a pioneering method for pinpointing PII-sensitive neurons (privacy neurons) within LLMs. Our method employs learnable binary weight masks to localize specific neurons that account for the memorization of PII in LLMs through adversarial training. Our investigations discover that PII is memorized by a small subset of neurons across all layers, which shows the property of PII specificity. Furthermore, we propose to validate the potential in PII risk mitigation by deactivating the localized privacy neurons. Both quantitative and qualitative experiments demonstrate the effectiveness of our neuron localization algorithm., Comment: ACL 2024 main conference

The interplay between nontrivial band topology and layered antiferromagnetism in MnBi2Te4 has opened up a new avenue for exploring topological phases of matter. Representative examples include the quantum anomalous Hall effect and axion insulator state observed in odd and even number layers of MnBi2Te4, when the top and bottom surfaces have parallel and antiparallel spin alignments respectively. The rich and complex spin dynamics associated with the van der Waals antiferromagnetic order is expected to generate novel topological phases and phase transitions that are unique to MnBi2Te4. Here we fabricate a device of 7-septuple-layer MnBi2Te4 covered with AlOx capping layer, which enables the investigation of antiferromagnetic quantum anomalous Hall effect over wide parameter spaces. By tuning the gate voltage and perpendicular magnetic field, we uncover a cascade of quantum phase transitions that can be attributed to the influence of spin configurations on charge transport. Furthermore, we find that an in-plane magnetic field enhances both the coercive field and exchange gap of the surface state, in sharp contrast to that in ferromagnetic quantum anomalous Hall state. We propose that these peculiar features arise from the spin flip and flop transitions inherent to van der Waals antiferromagnet. The versatile tunability of the quantum anomalous Hall effect in MnBi2Te4 paves the way for potential applications in topological antiferromagnetic spintronics., Comment: 16 pages, 4 figures

The rapidly developing deep learning (DL) techniques have been applied in software systems with various application scenarios. However, they could also pose new safety threats with potentially serious consequences, especially in safety-critical domains. DL libraries serve as the underlying foundation for DL systems, and bugs in them can have unpredictable impacts that directly affect the behaviors of DL systems. Previous research on fuzzing DL libraries still has limitations in the diversity of test inputs, the construction of test oracles, and the precision of detection. In this paper, we propose MoCo, a novel fuzzing testing method for DL libraries via assembling code. MoCo first disassembles the seed code file to obtain the template and code blocks, and then employs code block mutation operators (e.g., API replacement, random generation, and boundary checking) to generate more new code blocks adapted to the template. By inserting context-appropriate code blocks into the template step by step, MoCo can generate a tree of code files with intergenerational relations. According to the derivation relations in this tree and the applied mutation operators, we construct the test oracle based on the execution state consistency. Since the granularity of code assembly and mutation is controlled rather than randomly divergent, we can quickly pinpoint the lines of code where the bugs are located and the corresponding triggering conditions. We conduct a comprehensive experiment to evaluate the efficiency and effectiveness of MoCo using three widely-used DL libraries (i.e., TensorFlow, PyTorch, and Jittor). During the experiment, MoCo detects 64 new bugs of four types in three DL libraries, where 51 bugs have been confirmed, and 13 bugs have been fixed by developers.

Medical Visual Question Answering (MedVQA), which offers language responses to image-based medical inquiries, represents a challenging task and significant advancement in healthcare. It assists medical experts to swiftly interpret medical images, thereby enabling faster and more accurate diagnoses. However, the model interpretability and transparency of existing MedVQA solutions are often limited, posing challenges in understanding their decision-making processes. To address this issue, we devise a semi-automated annotation process to streamline data preparation and build new benchmark MedVQA datasets R-RAD, R-SLAKE and R-Path. These datasets provide intermediate medical decision-making rationales generated by multimodal large language models and human annotations for question-answering pairs in existing MedVQA datasets, i.e., VQA-RAD, SLAKE and PathVQA. Moreover, we design a novel framework, MedThink, which finetunes lightweight pretrained generative models by incorporating medical decision-making rationales. MedThink includes three distinct strategies to generate decision outcomes and corresponding rationales, thereby clearly showcasing the medical decision-making process during reasoning. Our comprehensive experiments show that our method achieves an accuracy of 83.5% on R-RAD, 86.3% on R-SLAKE and 87.2% on R-Path. These results significantly exceed those of existing state-of-the-art models with comparable parameters. Datasets and code will be released.

In the realm of aerospace design, achieving smooth curves is paramount, particularly when crafting objects such as airfoils. Generative Adversarial Network (GAN), a widely employed generative AI technique, has proven instrumental in synthesizing airfoil designs. However, a common limitation of GAN is the inherent lack of smoothness in the generated airfoil surfaces. To address this issue, we present a GAN model featuring a customized loss function built to produce seamlessly contoured airfoil designs. Additionally, our model demonstrates a substantial increase in design diversity compared to a conventional GAN augmented with a post-processing smoothing filter.

Anderson localization is anticipated to play a pivotal role in the manifestation of the quantum anomalous Hall effect, akin to its role in conventional quantum Hall effects. The significance of Anderson localization is particularly pronounced in elucidating the reasons behind the fragility of the observed quantum anomalous Hall state in the intrinsic magnetic topological insulator MnBi2Te4 with a large predicted magnetic gap. Here, employing varying sized MnBi2Te4 micro/nano-structures fabricated from a single molecular-beam-epitaxy-grown thin film, we have carried out a systematic size- and temperature-dependent study on the transport properties of the films regarding the quantum anomalous Hall states. The low-temperature transport properties of the finite-sized MnBi2Te4 samples can be quantitatively understood through Anderson localization, which plays an indispensable role in stabilizing the ground states. At higher temperatures, the failure of electron localization induced by an excessively short electronic dephasing length is identified as the cause of deviation from quantization. The work reveals that electronic dephasing and localization are non-negligible factors in designing high-temperature quantum anomalous Hall systems., Comment: 20 pages, 4 figures

Multimodal Large Language Models (MLLMs) such as GPT-4V and Gemini Pro face challenges in achieving human-level perception in Visual Question Answering (VQA), particularly in object-oriented perception tasks which demand fine-grained understanding of object identities, locations or attributes, as indicated by empirical findings. This is mainly due to their limited capability to effectively integrate complex visual cues with textual information and potential object hallucinations. In this paper, we present a novel approach, Joint Visual and Text Prompting (VTPrompt), that employs fine-grained visual information to enhance the capability of MLLMs in VQA, especially for object-oriented perception. VTPrompt merges visual and text prompts to extract key concepts from textual questions and employs a detection model to highlight relevant objects as visual prompts in images. The processed images alongside text prompts are subsequently fed into MLLMs to produce more accurate answers. Our experiments with GPT-4V and Gemini Pro, on three benchmarks, i.e., MME , MMB and POPE, demonstrate significant improvements. Particularly, our method led to a score improvement of up to 183.5 for GPT-4V on MME and enhanced MMB performance by 8.17\% for GPT-4V and 15.69\% for Gemini Pro.

Federated learning is gaining increasing popularity, with data heterogeneity and privacy being two prominent challenges. In this paper, we address both issues within a federated transfer learning framework, aiming to enhance learning on a target data set by leveraging information from multiple heterogeneous source data sets while adhering to privacy constraints. We rigorously formulate the notion of \textit{federated differential privacy}, which offers privacy guarantees for each data set without assuming a trusted central server. Under this privacy constraint, we study three classical statistical problems, namely univariate mean estimation, low-dimensional linear regression, and high-dimensional linear regression. By investigating the minimax rates and identifying the costs of privacy for these problems, we show that federated differential privacy is an intermediate privacy model between the well-established local and central models of differential privacy. Our analyses incorporate data heterogeneity and privacy, highlighting the fundamental costs of both in federated learning and underscoring the benefit of knowledge transfer across data sets., Comment: 78 pages, 3 figures

Although Large Language Models (LLMs) have demonstrated impressive text generation capabilities, they are easily misled by untruthful contexts provided by users or knowledge augmentation tools, leading to hallucinations. To alleviate LLMs from being misled by untruthful context and take advantage of knowledge augmentation, we propose Truth-Aware Context Selection (TACS), a lightweight method to adaptively recognize and mask untruthful context from the inputs. TACS begins by performing truth detection on the input context, leveraging the parameterized knowledge within the LLM. Subsequently, it constructs a corresponding attention mask based on the truthfulness of each position, selecting the truthful context and discarding the untruthful context. Additionally, we introduce a new evaluation metric, Disturbance Adaption Rate, to further study the LLMs' ability to accept truthful information and resist untruthful information. Experimental results indicate that TACS can effectively filter untruthful context and significantly improve the overall quality of LLMs' responses when presented with misleading information., Comment: Accepted to ACL 2024 Findings. Code is available at: https://github.com/ictnlp/TACS

Large Language Models (LLMs) sometimes suffer from producing hallucinations, especially LLMs may generate untruthful responses despite knowing the correct knowledge. Activating the truthfulness within LLM is the key to fully unlocking LLM's knowledge potential. In this paper, we propose TruthX, an inference-time intervention method to activate the truthfulness of LLM by identifying and editing the features within LLM's internal representations that govern the truthfulness. TruthX employs an auto-encoder to map LLM's representations into semantic and truthful latent spaces respectively, and applies contrastive learning to identify a truthful editing direction within the truthful space. During inference, by editing LLM's internal representations in truthful space, TruthX effectively enhances the truthfulness of LLM. Experiments show that TruthX improves the truthfulness of 13 advanced LLMs by an average of 20% on TruthfulQA benchmark. Further analyses suggest that TruthX can control LLM to produce truthful or hallucinatory responses via editing only one vector in LLM's internal representations., Comment: Accepted to ACL 2024 main conference, Project Page: https://ictnlp.github.io/TruthX-site/

Large Language Models (LLMs) have achieved impressive results in Machine Translation (MT). However, careful evaluations by human reveal that the translations produced by LLMs still contain multiple errors. Importantly, feeding back such error information into the LLMs can lead to self-refinement and result in improved translation performance. Motivated by these insights, we introduce a systematic LLM-based self-refinement translation framework, named \textbf{TEaR}, which stands for \textbf{T}ranslate, \textbf{E}stimate, \textbf{a}nd \textbf{R}efine, marking a significant step forward in this direction. Our findings demonstrate that 1) our self-refinement framework successfully assists LLMs in improving their translation quality across a wide range of languages, whether it's from high-resource languages to low-resource ones or whether it's English-centric or centered around other languages; 2) TEaR exhibits superior systematicity and interpretability; 3) different estimation strategies yield varied impacts, directly affecting the effectiveness of the final corrections. Additionally, traditional neural translation models and evaluation models operate separately, often focusing on singular tasks due to their limited capabilities, while general-purpose LLMs possess the capability to undertake both tasks simultaneously. We further conduct cross-model correction experiments to investigate the potential relationship between the translation and evaluation capabilities of general-purpose LLMs. Our code and data are available at https://github.com/fzp0424/self_correct_mt, Comment: Our code and data are available at https://github.com/fzp0424/self_correct_mt

Simultaneous Machine Translation (SiMT) generates translations while reading the source sentence, necessitating a policy to determine the optimal timing for reading and generating words. Despite the remarkable performance achieved by Large Language Models (LLM) across various NLP tasks, existing SiMT methods predominantly focus on conventional transformers, employing a single model to concurrently determine the policy and generate the translations. However, given the complexity of SiMT, it is challenging to effectively address both tasks with a single model. Therefore, there is a need to decouple the SiMT task into policy-decision and translation sub-tasks. We propose SiLLM, which delegates the two sub-tasks to separate agents, thereby incorporating LLM into SiMT. The policy-decision agent is managed by a conventional SiMT model, responsible for determining the translation policy. The translation agent, leveraging the capabilities of LLM, generates translation using the partial source sentence. The two agents collaborate to accomplish SiMT. To facilitate the application of token-level policies determined by conventional SiMT models to LLM, we propose a word-level policy adapted for LLM. Experiments on two datasets demonstrate that, with a small amount of data for fine-tuning LLM, SiLLM attains state-of-the-art performance., Comment: 13 pages, 6 tables, 7 figures

Task-oriented dialog systems have witnessed substantial progress due to conversational pre-training techniques. Yet, two significant challenges persist. First, most systems primarily utilize the latest turn's state label for the generator. This practice overlooks the comprehensive value of state labels in boosting the model's understanding for future generations. Second, an overreliance on generated policy often leads to error accumulation, resulting in suboptimal responses when adhering to incorrect actions. To combat these challenges, we propose turn-level multi-task objectives for the encoder. With the guidance of essential information from labeled intermediate states, we establish a more robust representation for both understanding and generation. For the decoder, we introduce an action tree-based scheduled sampling technique. Specifically, we model the hierarchical policy as trees and utilize the similarity between trees to sample negative policy based on scheduled sampling, hoping the model to generate invariant responses under perturbations. This method simulates potential pitfalls by sampling similar negative policy, bridging the gap between task-oriented dialog training and inference. Among methods without continual pre-training, our approach achieved state-of-the-art (SOTA) performance on the MultiWOZ dataset series and was also competitive with pre-trained SOTA methods., Comment: Accepted by AAAI 2024

Multi-sensor fusion stands as a pivotal technique in addressing numerous safety-critical tasks and applications, e.g., self-driving cars and automated robotic arms. With the continuous advancement in data-driven artificial intelligence (AI), MSF's potential for sensing and understanding intricate external environments has been further amplified, bringing a profound impact on intelligent systems and specifically on their perception systems. Similar to traditional software, adequate testing is also required for AI-enabled MSF systems. Yet, existing testing methods primarily concentrate on single-sensor perception systems (e.g., image-/point cloud-based object detection systems). There remains a lack of emphasis on generating multi-modal test cases for MSF systems. To address these limitations, we design and implement MultiTest, a fitness-guided metamorphic testing method for complex MSF perception systems. MultiTest employs a physical-aware approach to synthesize realistic multi-modal object instances and insert them into critical positions of background images and point clouds. A fitness metric is designed to guide and boost the test generation process. We conduct extensive experiments with five SOTA perception systems to evaluate MultiTest from the perspectives of: (1) generated test cases' realism, (2) fault detection capabilities, and (3) performance improvement. The results show that MultiTest can generate realistic and modality-consistent test data and effectively detect hundreds of diverse faults of an MSF system under test. Moreover, retraining an MSF system on the test cases generated by MultiTest can improve the system's robustness., Comment: The first two authors contributed equally. To appear in the proceedings of the 46th International Conference on Software Engineering (ICSE 2024)

The study of di-Higgs events, both resonant and non-resonant, plays a crucial role in understanding the fundamental interactions of the Higgs boson. In this work we consider di-Higgs events decaying into four $b$-quarks and propose to improve the experimental sensitivity by utilizing a novel machine learning algorithm known as Symmetry Preserving Attention Network (\textsc{Spa-Net}) -- a neural network structure whose architecture is designed to incorporate the inherent symmetries in particle reconstruction tasks. We demonstrate that the \textsc{Spa-Net} can enhance the experimental reach over baseline methods such as the cut-based and the Deep Neural Networks (DNN)-based analyses. At the Large Hadron Collider, with a 14-TeV centre-of-mass energy and an integrated luminosity of 300 fb$^{-1}$, the \textsc{Spa-Net} allows us to establish 95\% C.L. upper limits in resonant production cross-sections that are 10\% to 45\% stronger than baseline methods. For non-resonant di-Higgs production, \textsc{Spa-Net} enables us to constrain the self-coupling that is 9\% more stringent than the baseline method.

We use magnetic force microscopy (MFM) to study spatial uniformity of magnetization of epitaxially grown MnBi2Te4 thin films. Compared to films which exhibit no quantum anomalous Hall effect (QAH), films with QAH are observed to have more spatial uniformity of magnetization with larger domain size. The domain evolution upon magnetic field sweeping indicates that the magnetic domains or the spatial nonuniformity of magnetization originates from the strong pinning of the inherent sample inhomogeneity. A direct correlation between the Hall resistivity and the domain size has been established by analyzing a series of thin films with and without QAH. Our observation shows that one has to suppress the spatial nonuniformity of magnetization to allow the Hall resistivity to be quantized. The fact that a sizable longitudinal resistivity remains even for the QAH sample suggests a quantized Hall insulator scenario. Our work provides important insights to the understanding of the quantization mechanism and the dissipation of the QAH state in MnBi2Te4 system., Comment: 14 pages, 4 figures

In this study, we investigate intrinsic magnetic topological insulator MnBi2Te4 thin films grown by molecular beam epitaxy. We observe a reentrant quantum anomalous Hall effect when the Fermi energy enters the valance band and magnetic field equals zero, indicating the emergence of the Chern Anderson insulator state. The discovery opens a new avenue for realizing the QAH effect and underscores the fundamental role of both Berry curvature and Anderson localization., Comment: 15 pages, 4 figures

Federated Long-Tailed Learning (Fed-LT), a paradigm wherein data collected from decentralized local clients manifests a globally prevalent long-tailed distribution, has garnered considerable attention in recent times. In the context of Fed-LT, existing works have predominantly centered on addressing the data imbalance issue to enhance the efficacy of the generic global model while neglecting the performance at the local level. In contrast, conventional Personalized Federated Learning (pFL) techniques are primarily devised to optimize personalized local models under the presumption of a balanced global data distribution. This paper introduces an approach termed Federated Local and Generic Model Training in Fed-LT (FedLoGe), which enhances both local and generic model performance through the integration of representation learning and classifier alignment within a neural collapse framework. Our investigation reveals the feasibility of employing a shared backbone as a foundational framework for capturing overarching global trends, while concurrently employing individualized classifiers to encapsulate distinct refinements stemming from each client's local features. Building upon this discovery, we establish the Static Sparse Equiangular Tight Frame Classifier (SSE-C), inspired by neural collapse principles that naturally prune extraneous noisy features and foster the acquisition of potent data representations. Furthermore, leveraging insights from imbalance neural collapse's classifier norm patterns, we develop Global and Local Adaptive Feature Realignment (GLA-FR) via an auxiliary global classifier and personalized Euclidean norm transfer to align global features with client preferences. Extensive experimental results on CIFAR-10/100-LT, ImageNet, and iNaturalist demonstrate the advantage of our method over state-of-the-art pFL and Fed-LT approaches., Comment: Accepted by ICLR 2024, code: https://github.com/ZackZikaiXiao/FedLoGe

Nonlinear charge transport, such as nonreciprocal longitudinal resistance and nonlinear Hall effect, has attracted considerable interest in probing the symmetries and topological properties of new materials. Recent research has revealed significant nonreciprocal longitudinal resistance and nonlinear Hall effect in MnBi2Te4, an intrinsic magnetic topological insulator, induced by the quantum metric dipole. However, the inconsistent response with charge density and conflicting C3z symmetry requirement necessitate a thorough understanding of factors affecting the nonlinear transport measurement. This study uncovers an experimental factor leading to significant nonlinear transport signals in MnBi2Te4, attributed to gate voltage oscillation from the application of large alternating current. Additionally, a methodology is proposed to suppress this effect by individually grounding the voltage electrodes during second-harmonic measurements. The investigation underscores the critical importance of assessing the impact of gate voltage oscillation before determining the intrinsic nature of nonlinear transport in 2D material devices with an electrically connected operative gate electrode., Comment: 16 pages, 4 figures