arXiv subjects

Yun Zhu

Publications and source records attributed to Yun Zhu.

At least 73 records · Page 4Linked to original sources

arXiv2023

RewriteLM: An Instruction-Tuned Large Language Model for Text Rewriting

Large Language Models (LLMs) have demonstrated impressive capabilities in creative tasks such as storytelling and E-mail generation. However, as LLMs are primarily trained on final text results rather than intermediate revisions, it might be challenging for them to perform text rewriting tasks. Most studies in the rewriting tasks focus on a particular transformation type within the boundaries of single sentences. In this work, we develop new strategies for instruction tuning and reinforcement learning to better align LLMs for cross-sentence rewriting tasks using diverse wording and structures expressed through natural languages including 1) generating rewriting instruction data from Wiki edits and public corpus through instruction generation and chain-of-thought prompting; 2) collecting comparison data for reward model training through a new ranking function. To facilitate this research, we introduce OpenRewriteEval, a novel benchmark covers a wide variety of rewriting types expressed through natural language instructions. Our results show significant improvements over a variety of baselines. The public repository is available on GitHub under Google Research (https://github.com/google-research/google-research/tree/master/rewritelm).

cs.CL↗

arXiv2023

SiRA: Sparse Mixture of Low Rank Adaptation

Parameter Efficient Tuning has been an prominent approach to adapt the Large Language Model to downstream tasks. Most previous works considers adding the dense trainable parameters, where all parameters are used to adapt certain task. We found this less effective empirically using the example of LoRA that introducing more trainable parameters does not help. Motivated by this we investigate the importance of leveraging "sparse" computation and propose SiRA: sparse mixture of low rank adaption. SiRA leverages the Sparse Mixture of Expert(SMoE) to boost the performance of LoRA. Specifically it enforces the top $k$ experts routing with a capacity limit restricting the maximum number of tokens each expert can process. We propose a novel and simple expert dropout on top of gating network to reduce the over-fitting issue. Through extensive experiments, we verify SiRA performs better than LoRA and other mixture of expert approaches across different single tasks and multitask settings.

cs.CL↗

arXiv2023

Cappy: Outperforming and Boosting Large Multi-Task LMs with a Small Scorer

Large language models (LLMs) such as T0, FLAN, and OPT-IML, excel in multi-tasking under a unified instruction-following paradigm, where they also exhibit remarkable generalization abilities to unseen tasks. Despite their impressive performance, these LLMs, with sizes ranging from several billion to hundreds of billions of parameters, demand substantial computational resources, making their training and inference expensive and inefficient. Furthermore, adapting these models to downstream applications, particularly complex tasks, is often unfeasible due to the extensive hardware requirements for finetuning, even when utilizing parameter-efficient approaches such as prompt tuning. Additionally, the most powerful multi-task LLMs, such as OPT-IML-175B and FLAN-PaLM-540B, are not publicly accessible, severely limiting their customization potential. To address these challenges, we introduce a pretrained small scorer, Cappy, designed to enhance the performance and efficiency of multi-task LLMs. With merely 360 million parameters, Cappy functions either independently on classification tasks or serve as an auxiliary component for LLMs, boosting their performance. Moreover, Cappy enables efficiently integrating downstream supervision without requiring LLM finetuning nor the access to their parameters. Our experiments demonstrate that, when working independently on 11 language understanding tasks from PromptSource, Cappy outperforms LLMs that are several orders of magnitude larger. Besides, on 45 complex tasks from BIG-Bench, Cappy boosts the performance of the advanced multi-task LLM, FLAN-T5, by a large margin. Furthermore, Cappy is flexible to cooperate with other LLM adaptations, including finetuning and in-context learning, offering additional performance enhancement.

cs.LG↗

arXiv2023

Critique Ability of Large Language Models

Critical thinking is essential for rational decision-making and problem-solving. This skill hinges on the ability to provide precise and reasoned critiques and is a hallmark of human intelligence. In the era of large language models (LLMs), this study explores the ability of LLMs to deliver accurate critiques across various tasks. We are interested in this topic as a capable critic model could not only serve as a reliable evaluator, but also as a source of supervised signals for model tuning. Particularly, if a model can self-critique, it has the potential for autonomous self-improvement. To examine this, we introduce a unified evaluation framework for assessing the critique abilities of LLMs. We develop a benchmark called CriticBench, which comprises 3K high-quality natural language queries and corresponding model responses; and annotate the correctness of these responses. The benchmark cover tasks such as math problem-solving, code completion, and question answering. We evaluate multiple LLMs on the collected dataset and our analysis reveals several noteworthy insights: (1) Critique is generally challenging for most LLMs, and this capability often emerges only when models are sufficiently large. (2) In particular, self-critique is especially difficult. Even top-performing LLMs struggle to achieve satisfactory performance. (3) Models tend to have lower critique accuracy on problems where they are most uncertain. To this end, we introduce a simple yet effective baseline named self-check, which leverages self-critique to improve task performance for various models. We hope this study serves as an initial exploration into understanding the critique abilities of LLMs, and aims to inform future research, including the development of more proficient critic models and the application of critiques across diverse tasks.

cs.LG↗

arXiv2023

Cluster Language Model for Improved E-Commerce Retrieval and Ranking: Leveraging Query Similarity and Fine-Tuning for Personalized Results

This paper proposes a novel method to improve the accuracy of product search in e-commerce by utilizing a cluster language model. The method aims to address the limitations of the bi-encoder architecture while maintaining a minimal additional training burden. The approach involves labeling top products for each query, generating semantically similar query clusters using the K-Means clustering algorithm, and fine-tuning a global language model into cluster language models on individual clusters. The parameters of each cluster language model are fine-tuned to learn local manifolds in the feature space efficiently, capturing the nuances of various query types within each cluster. The inference is performed by assigning a new query to its respective cluster and utilizing the corresponding cluster language model for retrieval. The proposed method results in more accurate and personalized retrieval results, offering a superior alternative to the popular bi-encoder based retrieval models in semantic search.

cs.IR↗

arXiv2023

Towards an On-device Agent for Text Rewriting

Large Language Models (LLMs) have demonstrated impressive capabilities for text rewriting. Nonetheless, the large sizes of these models make them impractical for on-device inference, which would otherwise allow for enhanced privacy and economical inference. Creating a smaller yet potent language model for text rewriting presents a formidable challenge because it requires balancing the need for a small size with the need to retain the emergent capabilities of the LLM, that requires costly data collection. To address the above challenge, we introduce a new instruction tuning approach for building a mobile-centric text rewriting model. Our strategies enable the generation of high quality training data without any human labeling. In addition, we propose a heuristic reinforcement learning framework which substantially enhances performance without requiring preference data. To further bridge the performance gap with the larger server-side model, we propose an effective approach that combines the mobile rewrite agent with the server model using a cascade. To tailor the text rewriting tasks to mobile scenarios, we introduce MessageRewriteEval, a benchmark that focuses on text rewriting for messages through natural language instructions. Through empirical experiments, we demonstrate that our on-device model surpasses the current state-of-the-art LLMs in text rewriting while maintaining a significantly reduced model size. Notably, we show that our proposed cascading approach improves model performance.

cs.CL↗

arXiv2023

SGL-PT: A Strong Graph Learner with Graph Prompt Tuning

Recently, much exertion has been paid to design graph self-supervised methods to obtain generalized pre-trained models, and adapt pre-trained models onto downstream tasks through fine-tuning. However, there exists an inherent gap between pretext and downstream graph tasks, which insufficiently exerts the ability of pre-trained models and even leads to negative transfer. Meanwhile, prompt tuning has seen emerging success in natural language processing by aligning pre-training and fine-tuning with consistent training objectives. In this paper, we identify the challenges for graph prompt tuning: The first is the lack of a strong and universal pre-training task across sundry pre-training methods in graph domain. The second challenge lies in the difficulty of designing a consistent training objective for both pre-training and downstream tasks. To overcome above obstacles, we propose a novel framework named SGL-PT which follows the learning strategy ``Pre-train, Prompt, and Predict''. Specifically, we raise a strong and universal pre-training task coined as SGL that acquires the complementary merits of generative and contrastive self-supervised graph learning. And aiming for graph classification task, we unify pre-training and fine-tuning by designing a novel verbalizer-free prompting function, which reformulates the downstream task in a similar format as pretext task. Empirical results show that our method surpasses other baselines under unsupervised setting, and our prompt tuning method can greatly facilitate models on biological datasets over fine-tuning methods.

cs.LG↗

arXiv2023

MARIO: Model Agnostic Recipe for Improving OOD Generalization of Graph Contrastive Learning

In this work, we investigate the problem of out-of-distribution (OOD) generalization for unsupervised learning methods on graph data. This scenario is particularly challenging because graph neural networks (GNNs) have been shown to be sensitive to distributional shifts, even when labels are available. To address this challenge, we propose a \underline{M}odel-\underline{A}gnostic \underline{R}ecipe for \underline{I}mproving \underline{O}OD generalizability of unsupervised graph contrastive learning methods, which we refer to as MARIO. MARIO introduces two principles aimed at developing distributional-shift-robust graph contrastive methods to overcome the limitations of existing frameworks: (i) Information Bottleneck (IB) principle for achieving generalizable representations and (ii) Invariant principle that incorporates adversarial data augmentation to obtain invariant representations. To the best of our knowledge, this is the first work that investigates the OOD generalization problem of graph contrastive learning, with a specific focus on node-level tasks. Through extensive experiments, we demonstrate that our method achieves state-of-the-art performance on the OOD test set, while maintaining comparable performance on the in-distribution test set when compared to existing approaches. The source code for our method can be found at: https://github.com/ZhuYun97/MARIO

cs.LG↗

arXiv2023

Analysis and tuning of a three-term DMC

Most MPC (Model Predictive Control) algorithms used in industries and studied in the control academia use a two-term QP (quadratic programming), where the first term is the weighted norm of the output errors, and the second term is that of the input increments. In this work, a DMC (Dynamic Matrix Control) algorithm that uses three-term QP is studied, where the third term is the weighted norm of the output increments. In the analysis, a relationship between the three-term DMC and the two-term DMC is established; based on that, the closed-loop response curves are derived. Based on the analysis, two controller tuning procedures are developed for the three-term DMC, one for closed-loop step response and one for disturbance reduction. Finally, it will be proven that the three-term DMC can achieve a higher performance and robustness than the two-term DMC can. Simulation studies are used to demonstrate the findings and the tuning methods.

eess.SY↗

arXiv2023

SmartBERT: A Promotion of Dynamic Early Exiting Mechanism for Accelerating BERT Inference

Dynamic early exiting has been proven to improve the inference speed of the pre-trained language model like BERT. However, all samples must go through all consecutive layers before early exiting and more complex samples usually go through more layers, which still exists redundant computation. In this paper, we propose a novel dynamic early exiting combined with layer skipping for BERT inference named SmartBERT, which adds a skipping gate and an exiting operator into each layer of BERT. SmartBERT can adaptively skip some layers and adaptively choose whether to exit. Besides, we propose cross-layer contrastive learning and combine it into our training phases to boost the intermediate layers and classifiers which would be beneficial for early exiting. To keep the consistent usage of skipping gates between training and inference phases, we propose a hard weight mechanism during training phase. We conduct experiments on eight classification datasets of the GLUE benchmark. Experimental results show that SmartBERT achieves 2-3x computation reduction with minimal accuracy drops compared with BERT and our method outperforms previous methods in both efficiency and accuracy. Moreover, in some complex datasets like RTE and WNLI, we prove that the early exiting based on entropy hardly works, and the skipping mechanism is essential for reducing computation.

cs.CL↗

arXiv2023

Structure-Aware Group Discrimination with Adaptive-View Graph Encoder: A Fast Graph Contrastive Learning Framework

Albeit having gained significant progress lately, large-scale graph representation learning remains expensive to train and deploy for two main reasons: (i) the repetitive computation of multi-hop message passing and non-linearity in graph neural networks (GNNs); (ii) the computational cost of complex pairwise contrastive learning loss. Two main contributions are made in this paper targeting this twofold challenge: we first propose an adaptive-view graph neural encoder (AVGE) with a limited number of message passing to accelerate the forward pass computation, and then we propose a structure-aware group discrimination (SAGD) loss in our framework which avoids inefficient pairwise loss computing in most common GCL and improves the performance of the simple group discrimination. By the framework proposed, we manage to bring down the training and inference cost on various large-scale datasets by a significant margin (250x faster inference time) without loss of the downstream-task performance.

cs.LG↗

arXiv2022

Solving the Capsulation Attack against Backdoor-based Deep Neural Network Watermarks by Reversing Triggers

Backdoor-based watermarking schemes were proposed to protect the intellectual property of artificial intelligence models, especially deep neural networks, under the black-box setting. Compared with ordinary backdoors, backdoor-based watermarks need to digitally incorporate the owner's identity, which fact adds extra requirements to the trigger generation and verification programs. Moreover, these concerns produce additional security risks after the watermarking scheme has been published for as a forensics tool or the owner's evidence has been eavesdropped on. This paper proposes the capsulation attack, an efficient method that can invalidate most established backdoor-based watermarking schemes without sacrificing the pirated model's functionality. By encapsulating the deep neural network with a rule-based or Bayes filter, an adversary can block ownership probing and reject the ownership verification. We propose a metric, CAScore, to measure a backdoor-based watermarking scheme's security against the capsulation attack. This paper also proposes a new backdoor-based deep neural network watermarking scheme that is secure against the capsulation attack by reversing the encoding process and randomizing the exposure of triggers.

cs.CR↗

arXiv2022

RoSA: A Robust Self-Aligned Framework for Node-Node Graph Contrastive Learning

Graph contrastive learning has gained significant progress recently. However, existing works have rarely explored non-aligned node-node contrasting. In this paper, we propose a novel graph contrastive learning method named RoSA that focuses on utilizing non-aligned augmented views for node-level representation learning. First, we leverage the earth mover's distance to model the minimum effort to transform the distribution of one view to the other as our contrastive objective, which does not require alignment between views. Then we introduce adversarial training as an auxiliary method to increase sampling diversity and enhance the robustness of our model. Experimental results show that RoSA outperforms a series of graph contrastive learning frameworks on homophilous, non-homophilous and dynamic graphs, which validates the effectiveness of our work. To the best of our awareness, RoSA is the first work focuses on the non-aligned node-node graph contrastive learning problem. Our codes are available at: \href{https://github.com/ZhuYun97/RoSA}{\texttt{https://github.com/ZhuYun97/RoSA}}

cs.LG↗

arXiv2021

Fostering the Robustness of White-Box Deep Neural Network Watermarks by Neuron Alignment

The wide application of deep learning techniques is boosting the regulation of deep learning models, especially deep neural networks (DNN), as commercial products. A necessary prerequisite for such regulations is identifying the owner of deep neural networks, which is usually done through the watermark. Current DNN watermarking schemes, particularly white-box ones, are uniformly fragile against a family of functionality equivalence attacks, especially the neuron permutation. This operation can effortlessly invalidate the ownership proof and escape copyright regulations. To enhance the robustness of white-box DNN watermarking schemes, this paper presents a procedure that aligns neurons into the same order as when the watermark is embedded, so the watermark can be correctly recognized. This neuron alignment process significantly facilitates the functionality of established deep neural network watermarking schemes.

cs.CR↗

arXiv2021

Active Learning for Product Type Ontology Enhancement in E-commerce

Entity-based semantic search has been widely adopted in modern search engines to improve search accuracy by understanding users' intent. In e-commerce, an accurate and complete product type (PT) ontology is essential for recognizing product entities in queries and retrieving relevant products from catalog. However, finding product types (PTs) to construct such an ontology is usually expensive due to the considerable amount of human efforts it may involve. In this work, we propose an active learning framework that efficiently utilizes domain experts' knowledge for PT discovery. We also show the quality and coverage of the resulting PTs in the experiment results.

cs.LG↗

arXiv2020

Efficient Sensor Management for Multitarget Tracking in Passive Sensor Networks via Cauchy-Schwarz Divergence

This paper presents an efficient sensor management approach for multi-target tracking in passive sensor networks. Compared with active sensor networks, passive sensor networks have larger uncertainty due to the nature of passive sensing. Multi-target tracking in passive sensor networks is challenging because the multi-sensor multi-target fusion problem is difficult and sensor management is necessary to achieve good trade-offs between tracking accuracy and energy consumption or other costs. To address this problem, we present an efficient information-theoretic approach to manage the sensors for better tracking of the unknown and time-varying number of targets. This is accomplished with two main technical innovations. The first is a tractable information-based multi-sensor selection solution via a partially observed Markov decision process framework. The Cauchy-Schwarz divergence is used as the criterion to select informative sensors sequentially from the candidates. The second is a novel dual-stage fusion strategy based on the iterated-corrector multi-sensor generalized labeled multi-Bernoulli filter. Since the performance of the iterated-corrector scheme is greatly influenced by the order of sensor updates, the selected sensors are first ranked in order of their abilities to detect targets according to the Cauchy-Schwarz divergence, followed the iterated-corrector update. The computation costs of ranking the sensors are negligible, since the Cauchy-Schwarz divergence has been computed in the multi-sensor selection procedure. Simulation results validate the effectiveness and efficiency of the proposed approach.

eess.SP↗

arXiv2020

Fast Coherent Point Drift

Nonrigid point set registration is widely applied in the tasks of computer vision and pattern recognition. Coherent point drift (CPD) is a classical method for nonrigid point set registration. However, to solve spatial transformation functions, CPD has to compute inversion of a M*M matrix per iteration with time complexity O(M3). By introducing a simple corresponding constraint, we develop a fast implementation of CPD. The most advantage of our method is to avoid matrix-inverse operation. Before the iteration begins, our method requires to take eigenvalue decomposition of a M*M matrix once. After iteration begins, our method only needs to update a diagonal matrix with linear computational complexity, and perform matrix multiplication operation with time complexity approximately O(M2) in each iteration. Besides, our method can be further accelerated by the low-rank matrix approximation. Experimental results in 3D point cloud data show that our method can significantly reduce computation burden of the registration process, and keep comparable performance with CPD on accuracy.

cs.CV↗

arXiv2019

Shared Causal Paths underlying Alzheimer's dementia and Type 2 Diabetes

Background: Although Alzheimer's disease (AD) is a central nervous system disease and type 2 diabetes mellitus (T2DM) is a metabolic disorder, an increasing number of genetic epidemiological studies show clear link between AD and T2DM. The current approach to uncovering the shared pathways between AD and T2DM involves association analysis; however, such analyses lack power to discover the mechanisms of the diseases. Methods: We develop novel statistical methods to shift the current paradigm of genetic analysis from association analysis to deep causal inference for uncovering the shared mechanisms between AD and T2DM, and develop pipelines to infer multilevel omics causal networks which lead to shifting the current paradigm of genetic analysis from genetic analysis alone to integrated causal genomic, epigenomic, transcriptional and phenotypic data analysis. To discover common causal paths from genetic variants to AD and T2DM, we also develop algorithms that can automatically search the causal paths from genetic variants to diseases and Results: The proposed methods and algorithms are applied to ROSMAP dataset with 432 individuals who simultaneously had genotype, RNA-seq, DNA methylation and some phenotypes. We construct multi-omics causal networks and identify 13 shared causal genes, 16 shared causal pathways between AD and T2DM, and 754 gene expression and 101 gene methylation nodes that were connected to both AD and T2DM in multi-omics causal networks. Conclusions: The results of application of the proposed pipelines for identifying causal paths to real data analysis of AD and T2DM provided strong evidence to support the link between AD and T2DM and unraveled causal mechanism to explain this link.

q-bio.MN↗