Searcharxiv⌕ Search

arXiv subjects

Yi Zhou

Publications and source records attributed to Yi Zhou.

At least 307 records · Page 17Linked to original sources
arXiv2023

Forcing Generative Models to Degenerate Ones: The Power of Data Poisoning Attacks

Growing applications of large language models (LLMs) trained by a third party raise serious concerns on the security vulnerability of LLMs.It has been demonstrated that malicious actors can covertly exploit these vulnerabilities in LLMs through poisoning attacks aimed at generating undesirable outputs. While poisoning attacks have received significant attention in the image domain (e.g., object detection), and classification tasks, their implications for generative models, particularly in the realm of natural language generation (NLG) tasks, remain poorly understood. To bridge this gap, we perform a comprehensive exploration of various poisoning techniques to assess their effectiveness across a range of generative tasks. Furthermore, we introduce a range of metrics designed to quantify the success and stealthiness of poisoning attacks specifically tailored to NLG tasks. Through extensive experiments on multiple NLG tasks, LLMs and datasets, we show that it is possible to successfully poison an LLM during the fine-tuning stage using as little as 1\% of the total tuning data samples. Our paper presents the first systematic approach to comprehend poisoning attacks targeting NLG tasks considering a wide range of triggers and attack settings. We hope our findings will assist the AI security community in devising appropriate defenses against such threats.

cs.CR↗
arXiv2023

Segment Anything Model-guided Collaborative Learning Network for Scribble-supervised Polyp Segmentation

Polyp segmentation plays a vital role in accurately locating polyps at an early stage, which holds significant clinical importance for the prevention of colorectal cancer. Various polyp segmentation methods have been developed using fully-supervised deep learning techniques. However, pixel-wise annotation for polyp images by physicians during the diagnosis is both time-consuming and expensive. Moreover, visual foundation models such as the Segment Anything Model (SAM) have shown remarkable performance. Nevertheless, directly applying SAM to medical segmentation may not produce satisfactory results due to the inherent absence of medical knowledge. In this paper, we propose a novel SAM-guided Collaborative Learning Network (SAM-CLNet) for scribble-supervised polyp segmentation, enabling a collaborative learning process between our segmentation network and SAM to boost the model performance. Specifically, we first propose a Cross-level Enhancement and Aggregation Network (CEA-Net) for weakly-supervised polyp segmentation. Within CEA-Net, we propose a Cross-level Enhancement Module (CEM) that integrates the adjacent features to enhance the representation capabilities of different resolution features. Additionally, a Feature Aggregation Module (FAM) is employed to capture richer features across multiple levels. Moreover, we present a box-augmentation strategy that combines the segmentation maps generated by CEA-Net with scribble annotations to create more precise prompts. These prompts are then fed into SAM, generating segmentation SAM-guided masks, which can provide additional supervision to train CEA-Net effectively. Furthermore, we present an Image-level Filtering Mechanism to filter out unreliable SAM-guided masks. Extensive experimental results show that our SAM-CLNet outperforms state-of-the-art weakly-supervised segmentation methods.

cs.CV↗
arXiv2023

Fast Maximum $k$-Plex Algorithms Parameterized by Small Degeneracy Gaps

Given a graph, a $k$-plex is a set of vertices in which each vertex is not adjacent to at most $k-1$ other vertices in the set. The maximum $k$-plex problem, which asks for the largest $k$-plex from the given graph, is an important but computationally challenging problem in applications such as graph mining and community detection. So far, there are many practical algorithms, but without providing theoretical explanations on their efficiency. We define a novel parameter of the input instance, $g_k(G)$, the gap between the degeneracy bound and the size of the maximum $k$-plex in the given graph, and present an exact algorithm parameterized by this $g_k(G)$, which has a worst-case running time polynomial in the size of the input graph and exponential in $g_k(G)$. In real-world inputs, $g_k(G)$ is very small, usually bounded by $O(\log{(|V|)})$, indicating that the algorithm runs in polynomial time. We further extend our discussion to an even smaller parameter $cg_k(G)$, the gap between the community-degeneracy bound and the size of the maximum $k$-plex, and show that without much modification, our algorithm can also be parameterized by $cg_k(G)$. To verify the empirical performance of these algorithms, we carry out extensive experiments to show that these algorithms are competitive with the state-of-the-art algorithms. In particular, for large $k$ values such as $15$ and $20$, our algorithms dominate the existing algorithms. Finally, empirical analysis is performed to illustrate the effectiveness of the parameters and other key components in the implementation.

cs.DS↗
arXiv2023

Conic10K: A Challenging Math Problem Understanding and Reasoning Dataset

Mathematical understanding and reasoning are crucial tasks for assessing the capabilities of artificial intelligence (AI). However, existing benchmarks either require just a few steps of reasoning, or only contain a small amount of data in one specific topic, making it hard to analyse AI's behaviour with reference to different problems within a specific topic in detail. In this work, we propose Conic10K, a challenging math problem dataset on conic sections in Chinese senior high school education. Our dataset contains various problems with different reasoning depths, while only the knowledge from conic sections is required. Since the dataset only involves a narrow range of knowledge, it is easy to separately analyse the knowledge a model possesses and the reasoning ability it has. For each problem, we provide a high-quality formal representation, the reasoning steps, and the final solution. Experiments show that existing large language models, including GPT-4, exhibit weak performance on complex reasoning. We hope that our findings could inspire more advanced techniques for precise natural language understanding and reasoning. Our dataset and codes are available at https://github.com/whyNLP/Conic10K.

cs.CL↗
arXiv2023

A variational Monte Carlo approach to the SU(4) spin-orbital model on the triangular lattice

Previous investigations have suggested that the simplest spin-orbital model on the simplest frustrated lattice can host a nematic quantum spin-orbital liquid state. Namely, the orbital degeneracy of the SU(4) Kugel-Khomskii (KK) model tends to enhance quantum fluctuations and stabilize a quantum spin-orbital liquid exhibiting stripy features on the triangular lattice, as revealed by the state-of-the-art method of the density matrix renormalization group boosted by Gutzwiller projected wave functions. In this work, using the variational quantum Monte Carlo method, we have studied several spin-orbital liquid states, including a uniform $π$ flux state, three stripy states, and a plaquette state, on the $L\times{}L$ torus up to $L=24$. It turns out that one of these stripy states, called the "stripe-II" state, is energetically favored. This ground state breaks the $C_6$ symmetry of the lattice, resulting in a reduced $C_2$ symmetry and doubled unit cells, while preserving the SU(4) spin-orbital rotation symmetry. Such a nematic quantum spin-orbital liquid state can be characterized by a parton Fermi surface (FS) consisting of open orbits in the Brillouin zone, in contrast to the circular FS of the uniform $π$-flux state.

cond-mat.str-el↗
arXiv2023

Privacy-Preserving Federated Learning over Vertically and Horizontally Partitioned Data for Financial Anomaly Detection

The effective detection of evidence of financial anomalies requires collaboration among multiple entities who own a diverse set of data, such as a payment network system (PNS) and its partner banks. Trust among these financial institutions is limited by regulation and competition. Federated learning (FL) enables entities to collaboratively train a model when data is either vertically or horizontally partitioned across the entities. However, in real-world financial anomaly detection scenarios, the data is partitioned both vertically and horizontally and hence it is not possible to use existing FL approaches in a plug-and-play manner. Our novel solution, PV4FAD, combines fully homomorphic encryption (HE), secure multi-party computation (SMPC), differential privacy (DP), and randomization techniques to balance privacy and accuracy during training and to prevent inference threats at model deployment time. Our solution provides input privacy through HE and SMPC, and output privacy against inference time attacks through DP. Specifically, we show that, in the honest-but-curious threat model, banks do not learn any sensitive features about PNS transactions, and the PNS does not learn any information about the banks' dataset but only learns prediction labels. We also develop and analyze a DP mechanism to protect output privacy during inference. Our solution generates high-utility models by significantly reducing the per-bank noise level while satisfying distributed DP. To ensure high accuracy, our approach produces an ensemble model, in particular, a random forest. This enables us to take advantage of the well-known properties of ensembles to reduce variance and increase accuracy. Our solution won second prize in the first phase of the U.S. Privacy Enhancing Technologies (PETs) Prize Challenge.

cs.CR↗
arXiv2023

A Predictive Factor Analysis of Social Biases and Task-Performance in Pretrained Masked Language Models

Various types of social biases have been reported with pretrained Masked Language Models (MLMs) in prior work. However, multiple underlying factors are associated with an MLM such as its model size, size of the training data, training objectives, the domain from which pretraining data is sampled, tokenization, and languages present in the pretrained corpora, to name a few. It remains unclear as to which of those factors influence social biases that are learned by MLMs. To study the relationship between model factors and the social biases learned by an MLM, as well as the downstream task performance of the model, we conduct a comprehensive study over 39 pretrained MLMs covering different model sizes, training objectives, tokenization methods, training data domains and languages. Our results shed light on important factors often neglected in prior literature, such as tokenization or model objectives.

cs.CL↗
arXiv2023

An Efficient Multilingual Language Model Compression through Vocabulary Trimming

Multilingual language model (LM) have become a powerful tool in NLP especially for non-English languages. Nevertheless, model parameters of multilingual LMs remain large due to the larger embedding matrix of the vocabulary covering tokens in different languages. On the contrary, monolingual LMs can be trained in a target language with the language-specific vocabulary only, but this requires a large budget and availability of reliable corpora to achieve a high-quality LM from scratch. In this paper, we propose vocabulary-trimming (VT), a method to reduce a multilingual LM vocabulary to a target language by deleting irrelevant tokens from its vocabulary. In theory, VT can compress any existing multilingual LM to build monolingual LMs in any language covered by the multilingual LM. In our experiments, we show that VT can retain the original performance of the multilingual LM, while being smaller in size (in general around 50% of the original vocabulary size is enough) than the original multilingual LM. The evaluation is performed over four NLP tasks (two generative and two classification tasks) among four widely used multilingual LMs in seven languages. Finally, we show that this methodology can keep the best of both monolingual and multilingual worlds by keeping a small size as monolingual models without the need for specifically retraining them, and even limiting potentially harmful social biases.

cs.CL↗
arXiv2023

Can Word Sense Distribution Detect Semantic Changes of Words?

Semantic Change Detection (SCD) of words is an important task for various NLP applications that must make time-sensitive predictions. Some words are used over time in novel ways to express new meanings, and these new meanings establish themselves as novel senses of existing words. On the other hand, Word Sense Disambiguation (WSD) methods associate ambiguous words with sense ids, depending on the context in which they occur. Given this relationship between WSD and SCD, we explore the possibility of predicting whether a target word has its meaning changed between two corpora collected at different time steps, by comparing the distributions of senses of that word in each corpora. For this purpose, we use pretrained static sense embeddings to automatically annotate each occurrence of the target word in a corpus with a sense id. Next, we compute the distribution of sense ids of a target word in a given corpus. Finally, we use different divergence or distance measures to quantify the semantic change of the target word across the two given corpora. Our experimental results on SemEval 2020 Task 1 dataset show that word sense distributions can be accurately used to predict semantic changes of words in English, German, Swedish and Latin.

cs.CL↗
arXiv2023

Super Denoise Net: Speech Super Resolution with Noise Cancellation in Low Sampling Rate Noisy Environments

Speech super-resolution (SSR) aims to predict a high resolution (HR) speech signal from its low resolution (LR) corresponding part. Most neural SSR models focus on producing the final result in a noise-free environment by recovering the spectrogram of high-frequency part of the signal and concatenating it with the original low-frequency part. Although these methods achieve high accuracy, they become less effective when facing the real-world scenario, where unavoidable noise is present. To address this problem, we propose a Super Denoise Net (SDNet), a neural network for a joint task of super-resolution and noise reduction from a low sampling rate signal. To that end, we design gated convolution and lattice convolution blocks to enhance the repair capability and capture information in the time-frequency axis, respectively. The experiments show our method outperforms baseline speech denoising and SSR models on DNS 2020 no-reverb test set with higher objective and subjective scores.

eess.AS↗
arXiv2023

Global well-posedness for radial extremal hypersurface equation in $\left(1+3 \right)$-dimensional Minkowski space-time in critical Sobolev space

In this article, we prove the global well-posedness in the critical Sobolev space $H_{rad}^2\left(\mathbb{R}^2\right) \times H_{rad}^1 \left(\mathbb{R}^2\right)$ for the radial time-like extremal hypersurface equation in $\left(1+3\right)$- dimensional Minkowski space-time. This is achieved by deriving a new div-curl type lemma and combined it with energy and ``momentum" balance law to get some space-time estimates of the nonlinearity.

math.AP↗
arXiv2023

Uncertainty-inspired Open Set Learning for Retinal Anomaly Identification

Failure to recognize samples from the classes unseen during training is a major limitation of artificial intelligence in the real-world implementation for recognition and classification of retinal anomalies. We established an uncertainty-inspired open-set (UIOS) model, which was trained with fundus images of 9 retinal conditions. Besides assessing the probability of each category, UIOS also calculated an uncertainty score to express its confidence. Our UIOS model with thresholding strategy achieved an F1 score of 99.55%, 97.01% and 91.91% for the internal testing set, external target categories (TC)-JSIEC dataset and TC-unseen testing set, respectively, compared to the F1 score of 92.20%, 80.69% and 64.74% by the standard AI model. Furthermore, UIOS correctly predicted high uncertainty scores, which would prompt the need for a manual check in the datasets of non-target categories retinal diseases, low-quality fundus images, and non-fundus images. UIOS provides a robust method for real-world screening of retinal anomalies.

cs.LG↗
arXiv2023

Generalizable and explainable prediction of potential miRNA-disease associations based on heterogeneous graph learning

Biomedical research has revealed the crucial role of miRNAs in the progression of many diseases, and computational prediction methods are increasingly proposed for assisting biological experiments to verify miRNA-disease associations (MDAs). However, the generalizability and explainability are currently underemphasized. It's significant to generalize effective predictions to entities with fewer or no existing MDAs and reveal how the prediction scores are derived. In this study, our work contributes to data, model, and result analysis. First, for better formulation of the MDA issue, we integrate multi-source data into a heterogeneous graph with a broader learning and prediction scope, and we split massive verified MDAs into independent training, validation, and test sets as a benchmark. Second, we construct an end-to-end data-driven model that performs node feature encoding, graph structure learning, and binary prediction sequentially, with a heterogeneous graph transformer as the central module. Finally, computational experiments illustrate that our method outperforms existing state-of-the-art methods, achieving better evaluation metrics and alleviating the neglect of unknown miRNAs and diseases effectively. Case studies further demonstrate that we can make reliable MDA detections on diseases without MDA records, and the predictions can be explained in general and case by case.

cs.CE↗
arXiv2023

Tree-of-Mixed-Thought: Combining Fast and Slow Thinking for Multi-hop Visual Reasoning

There emerges a promising trend of using large language models (LLMs) to generate code-like plans for complex inference tasks such as visual reasoning. This paradigm, known as LLM-based planning, provides flexibility in problem solving and endows better interpretability. However, current research is mostly limited to basic scenarios of simple questions that can be straightforward answered in a few inference steps. Planning for the more challenging multi-hop visual reasoning tasks remains under-explored. Specifically, under multi-hop reasoning situations, the trade-off between accuracy and the complexity of plan-searching becomes prominent. The prevailing algorithms either address the efficiency issue by employing the fast one-stop generation or adopt a complex iterative generation method to improve accuracy. Both fail to balance the need for efficiency and performance. Drawing inspiration from the dual system of cognition in the human brain, the fast and the slow think processes, we propose a hierarchical plan-searching algorithm that integrates the one-stop reasoning (fast) and the Tree-of-thought (slow). Our approach succeeds in performance while significantly saving inference steps. Moreover, we repurpose the PTR and the CLEVER datasets, developing a systematic framework for evaluating the performance and efficiency of LLMs-based plan-search algorithms under reasoning tasks at different levels of difficulty. Extensive experiments demonstrate the superiority of our proposed algorithm in terms of performance and efficiency. The dataset and code will be release soon.

cs.CL↗
arXiv2023

Boosting Multi-modal Model Performance with Adaptive Gradient Modulation

While the field of multi-modal learning keeps growing fast, the deficiency of the standard joint training paradigm has become clear through recent studies. They attribute the sub-optimal performance of the jointly trained model to the modality competition phenomenon. Existing works attempt to improve the jointly trained model by modulating the training process. Despite their effectiveness, those methods can only apply to late fusion models. More importantly, the mechanism of the modality competition remains unexplored. In this paper, we first propose an adaptive gradient modulation method that can boost the performance of multi-modal models with various fusion strategies. Extensive experiments show that our method surpasses all existing modulation methods. Furthermore, to have a quantitative understanding of the modality competition and the mechanism behind the effectiveness of our modulation method, we introduce a novel metric to measure the competition strength. This metric is built on the mono-modal concept, a function that is designed to represent the competition-less state of a modality. Through systematic investigation, our results confirm the intuition that the modulation encourages the model to rely on the more informative modality. In addition, we find that the jointly trained model typically has a preferred modality on which the competition is weaker than other modalities. However, this preferred modality need not dominate others. Our code will be available at https://github.com/lihong2303/AGM_ICCV2023.

cs.CV↗
arXiv2023

Electromagnetic Spatiotemporal Differentiators

Spatiotemporal optical computing devices which could perform mathematical operations in both spatial and temporal domains can provide unprecedented measures to build efficient and real-time information processing systems. It is particularly important to realize the comprehensive functions in a compact design for better integration with electronic components. In this work, we experimentally demonstrated an analogue spatiotemporal differentiator in microwaves based on an asymmetrical metasurface which has a phase singularity in the spatiotemporal domain. We showed that this structure could give rise to a spatiotemporal transfer function required by an ideal first-order differentiator in both spatial and temporal domains by tailoring the unidirectional excitation of spoof surface plasmon polaritons (SSPPs). The spatial edge detection was performed utilizing a metallic slit and the temporal differentiation capability of the device was examined by Gaussian-like temporal pulses of different width. We further confirmed the differentiator demonstrated here could detect sharp changes of spatiotemporal pulses even with intricate profiles and theoretically estimated the resolution limits of the spatial and temporal edge detection. We also show that the pulse input after passing the spatiotemporal differentiator implemented here could carry a transverse orbital angular momentum (OAM) with a fractal topology charge which further increases the information quantity.

physics.app-ph↗