Query Feedback Research Articles

DifAttack: Query-Efficient Black-Box Adversarial Attack via Disentangled Feature Space

This work investigates efficient score-based black-box adversarial attacks with high Attack Success Rate (ASR) and good generalizability. We design a novel attack method based on a Disentangled Feature space, called DifAttack, which differs significantly from the existing ones operating over the entire feature space. Specifically, DifAttack firstly disentangles an image's latent feature into an adversarial feature and a visual feature, where the former dominates the adversarial capability of an image, while the latter largely determines its visual appearance. We train an autoencoder for the disentanglement by using pairs of clean images and their Adversarial Examples (AEs) generated from available surrogate models via white-box attack methods. Eventually, DifAttack iteratively optimizes the adversarial feature according to the query feedback from the victim model until a successful AE is generated, while keeping the visual feature unaltered. In addition, due to the avoidance of using surrogate models' gradient information when optimizing AEs for black-box models, our proposed DifAttack inherently possesses better attack capability in the open-set scenario, where the training dataset of the victim model is unknown. Extensive experimental results demonstrate that our method achieves significant improvements in ASR and query efficiency simultaneously, especially in the targeted attack and open-set scenarios. The code is available The code is available at https://github.com/csjunjun/DifAttack.git.

Generalizable Black-Box Adversarial Attack With Meta Learning.

In the scenario of black-box adversarial attack, the target model's parameters are unknown, and the attacker aims to find a successful adversarial perturbation based on query feedback under a query budget. Due to the limited feedback information, existing query-based black-box attack methods often require many queries for attacking each benign example. To reduce query cost, we propose to utilize the feedback information across historical attacks, dubbed example-level adversarial transferability. Specifically, by treating the attack on each benign example as one task, we develop a meta-learning framework by training a meta generator to produce perturbations conditioned on benign examples. When attacking a new benign example, the meta generator can be quickly fine-tuned based on the feedback information of the new task as well as a few historical attacks to produce effective perturbations. Moreover, since the meta-train procedure consumes many queries to learn a generalizable generator, we utilize model-level adversarial transferability to train the meta generator on a white-box surrogate model, then transfer it to help the attack against the target model. The proposed framework with the two types of adversarial transferability can be naturally combined with any off-the-shelf query-based attack methods to boost their performance, which is verified by extensive experiments. The source code is available at https://github.com/SCLBD/MCG-Blackbox.

Query Attack by Multi-Identity Surrogates

Deep Neural Networks (DNNs) are acknowledged as vulnerable to adversarial attacks, while the existing black-box attacks require extensive queries on the victim DNN to achieve high success rates. For query-efficiency, surrogate models of the victim are used to generate transferable Adversarial Examples (AEs) because of their <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Gradient Similarity (GS)</i> , i.e., surrogates' attack gradients are similar to the victim's ones. However, it is generally neglected to exploit their similarity on outputs, namely the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Prediction Similarity (PS)</i> , to filter out inefficient queries by surrogates without querying the victim. To jointly utilize and also optimize surrogates' GS and PS, we develop QueryNet, a unified attack framework that can significantly reduce queries. QueryNet creatively attacks by multi-identity surrogates, i.e., crafts several AEs for one sample by different surrogates, and also uses surrogates to decide on the most promising AE for the query. After that, the victim's query feedback is accumulated to optimize not only surrogates' parameters but also their architectures, enhancing both the GS and the PS. Although QueryNet has no access to pre-trained surrogates' prior, it reduces queries by averagely about an order of magnitude compared to alternatives within an acceptable time, according to our comprehensive experiments: 11 victims (including two commercial models) on MNIST/CIFAR10/ImageNet, allowing only 8-bit image queries, and no access to the victim's training data. The code is available at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/Sizhe-Chen/QueryNet</uri> .

An Embeddable Implicit IUVD Representation for Part-Based 3D Human Surface Reconstruction.

To reconstruct a 3D human surface from a single image, it is crucial to simultaneously consider human pose, shape, and clothing details. Recent approaches have combined parametric body models (such as SMPL), which capture body pose and shape priors, with neural implicit functions that flexibly learn clothing details. However, this combined representation introduces additional computation, e.g. signed distance calculation in 3D body feature extraction, leading to redundancy in the implicit query-and-infer process and failing to preserve the underlying body shape prior. To address these issues, we propose a novel IUVD-Feedback representation, consisting of an IUVD occupancy function and a feedback query algorithm. This representation replaces the time-consuming signed distance calculation with a simple linear transformation in the IUVD space, leveraging the SMPL UV maps. Additionally, it reduces redundant query points through a feedback mechanism, leading to more reasonable 3D body features and more effective query points, thereby preserving the parametric body prior. Moreover, the IUVD-Feedback representation can be embedded into any existing implicit human reconstruction pipeline without requiring modifications to the trained neural networks. Experiments on the THuman2.0 dataset demonstrate that the proposed IUVD-Feedback representation improves the robustness of results and achieves three times faster acceleration in the query-and-infer process. Furthermore, this representation holds potential for generative applications by leveraging its inherent semantic information from the parametric body model.

DDSG-GAN: Generative Adversarial Network with Dual Discriminators and Single Generator for Black-Box Attacks

As one of the top ten security threats faced by artificial intelligence, the adversarial attack has caused scholars to think deeply from theory to practice. However, in the black-box attack scenario, how to raise the visual quality of an adversarial example (AE) and perform a more efficient query should be further explored. This study aims to use the architecture of GAN combined with the model-stealing attack to train surrogate models and generate high-quality AE. This study proposes an image AE generation method based on the generative adversarial networks with dual discriminators and a single generator (DDSG-GAN) and designs the corresponding loss function for each model. The generator can generate adversarial perturbation, and two discriminators constrain the perturbation, respectively, to ensure the visual quality and attack effect of the generated AE. We extensively experiment on MNIST, CIFAR10, and Tiny-ImageNet datasets. The experimental results illustrate that our method can effectively use query feedback to generate an AE, which significantly reduces the number of queries on the target model and can implement effective attacks.

Turbo-charging SPJ query plans with learned physical join operator selections

The optimization of select-project-join (SPJ) queries entails two major challenges: (i) finding a good join order and (ii) selecting the best-fitting physical join operator for each single join within the chosen join order. Previous work mainly focuses on the computation of a good join order, but leaves open to which extent the physical join operator selection accounts for plan quality. Our analysis using different query optimizers indicates that physical join operator selection is crucial and that none of the investigated query optimizers reaches the full potential of optimal operator selections. To unlock this potential, we propose TONIC , a novel cardinality estimation-free extension for generic SPJ query optimizers in this paper. TONIC follows a learning-based approach and revises operator decisions for arbitrary join paths based on learned query feedback. To continuously capture and reuse optimal operator selections, we introduce a lightweight yet powerful Query Execution Plan Synopsis ( QEP-S ). In comparison to related work, TONIC enables transparent planning decisions with consistent performance improvements. Using two real-life benchmarks, we demonstrate that extending existing optimizers with TONIC substantially reduces query response times with a cumulative speedup of up to 2.8x.

Human-Feedback Shield Synthesis for Perceived Safety in Deep Reinforcement Learning

Despite the successes of deep reinforcement learning (RL), it is still challenging to obtain safe policies. Formal verification approaches ensure safety at all times, but usually overly restrict the agent's behaviors, since they assume adversarial behavior of the environment. Instead of assuming adversarial behavior, we suggest to focus on perceived safety instead, i.e., policies that avoid undesired behaviors while having a desired level of conservativeness. To obtain policies that are perceived as safe, we propose a shield synthesis framework with two distinct loops: (1) an inner loop that trains a policy with a set of actions that is constrained by shields whose conservativeness is parameterized, and (2) an outer loop that presents example rollouts of the policy to humans and collects their feedback to update the parameters of the shields in the inner loop. We demonstrate our approach on a RL benchmark of Lunar landing and a scenario in which a mobile robot navigates around humans. For the latter, we conducted two user studies to obtain policies that were perceived as safe. Our results indicate that our framework converges to policies that are perceived as safe, is robust against noisy feedback, and can query feedback for multiple policies at the same time.

Blockchain technology for immunisation documentation in India: findings from a simulation pilot

BackgroundImmunisation documentation in India has historically been fragmented over a number of systems. To reduce errors and wasted vaccinations for one of the largest and most complex public vaccination systems, data storage has migrated to electronic application-based data records. However, the system remains fragmented. There has been much interest in the global health sector about the use of blockchain technology to build secure, immutable databases for complex data in fragmented environments. As part of the Indian government's innovation initiatives, our project aims to describe a blockchain immunisation system to store immunisation data, pilot its use in India, and assess strengths, limitations, and feasibility of blockchain immunisation. MethodsA connector database—Anveshan—was developed for the state of Gujarat, India, to: (1) remove data fragmentation; (2) enable aggregate analysis to allow tracking of dynamic data; (3) track data changes; and (4) enable fault tolerance so data can be accurately preserved for long periods of time. Although live piloting was not possible because of the COVID-19 pandemic, a simulated pilot was done. A report including lessons learned, benefits and challenges of blockchain application to immunisation documentation, and recommendations for future usage was delivered. FindingsResults from the simulation showed that the Anveshan database was able to decrease data fragmentation by querying across both supply and delivery databases. Aggregate analysis capability was increased through development of a dashboard user interface that allowed policymakers and immunisation programme staff to receive real-time query feedback. Data changes could be tracked and private data encrypted by incorporating disincentives for inappropriate data modification. Blockchain technology properties increased fault tolerance, which make data less corruptible. Although blockchain technology offers benefits for data storage, it does not facilitate improvements in data collection and quality, which remain a significant challenge in India. Multiple health-care stakeholders would need to work together and have access to live data to ensure nationwide scalability. InterpretationOur findings show the challenges and benefits of a private blockchain system and identify areas requiring continued improvement for immunisation documentation in India. One important consideration in infrastructure design on a nationwide scale is the design of the blockchain system itself. This study is an important first step to show that evidence-based design innovation using blockchain can help countries such as India to address complex public health-care delivery conundrums. FundingIndia Grand Challenges grant (Bill & Melinda Gates Foundation and BIRAC, Government of India).

Incorporating Semantic Word Representations into Query Expansion for Microblog Information Retrieval

Microblog information retrieval has attracted much attention of researchers to capture the desired information in daily communications on social networks. Since the contents of microblogs are always non-standardized and flexible, including many popular Internet expressions, the retrieval accuracy of microblogs has much room for improvement. To enhance microblog information retrieval, we propose a novel query expansion method to enrich user queries with semantic word representations. In our method, we use a neural network model to map each word in the corpus to a low-dimensional vector representation. The mapped word vectors satisfy the algebraic vector addition operation, and the new vector obtained by the addition operation can express some common attributes of the two words. In this sense, we represent keywords in user queries as vectors, sum all the keyword vectors, and use the obtained query vectors to select the expansion words. In addition, we also combine the traditional pseudo-relevance feedback query expansion method with the proposed query expansion method. Experimental results show that the proposed method is effective and reduces noises in the expanded query, which improves the accuracy of microblog retrieval.

Real-time feedback query expansion technique for supporting scholarly search using citation network analysis

Scholars routinely search relevant papers to discover and put a new idea into proper context. Despite ongoing advances in scholarly retrieval technologies, locating relevant papers through keyword queries is still quite challenging due to the massive expansion in the size of the research paper repository. To tackle this problem, we propose a novel real-time feedback query expansion technique, which is a two-stage interactive scholarly search process. Upon receiving the initial search query, the retrieval system provides a ranked list of results. In the second stage, a user selects a few relevant papers, from which useful terms are extracted for query expansion. The newly expanded query is run against the index in real time to generate the final list of research papers. In both stages, citation analysis is involved in further improving the quality of the results. The novelty of the approach lies in the combined exploitation of query expansion and citation analysis that may bring the most relevant papers to the top of the search results list. The experimental results on the Association of Computational Linguistics (ACL) Anthology Network data set demonstrate that this technique is effective and robust for locating relevant papers regarding normalised discounted cumulative gain (nDCG), precision and recall rates than several state-of-the-art approaches.

Verbosity normalized pseudo-relevance feedback in information retrieval

Document length normalization is one of the fundamental components in a retrieval model because term frequencies can readily be increased in long documents. The key hypotheses in literature regarding document length normalization are the verbosity and scope hypotheses, which imply that document length normalization should consider the distinguishing effects of verbosity and scope on term frequencies. In this article, we extend these hypotheses in a pseudo-relevance feedback setting by assuming the verbosity hypothesis on the feedback query model, which states that the verbosity of an expanded query should not be high. Furthermore, we postulate the following two effects of document verbosity on a feedback query model that easily and typically holds in modern pseudo-relevance feedback methods: 1) the verbosity-preserving effect: the query verbosity of a feedback query model is determined by feedback document verbosities; 2) the verbosity-sensitive effect: highly verbose documents more significantly and unfairly affect the resulting query model than normal documents do. By considering these effects, we propose verbosity normalized pseudo-relevance feedback, which is straightforwardly obtained by replacing original term frequencies with their verbosity-normalized term frequencies in the pseudo-relevance feedback method. The results of the experiments performed on three standard TREC collections show that the proposed verbosity normalized pseudo-relevance feedback consistently provides statistically significant improvements over conventional methods, under the settings of the relevance model and latent concept expansion.

A Cardinality Estimation Approach Based on Two Level Histograms

For the mainstream relational database management systems, histograms play important roles in cardinality estimation. The main histogram-based cardinality estimation approaches can be classified into two categories: proactive approaches and reactive approaches. For the former, histograms are constructed and updated by periodical data scan which is also the essential reason affecting the accuracy and performance of this kind of approaches. Data scan is avoided in the latter, as an alternative, query feedback records (QFRs) are collected to construct and update histograms. But some time-con- suming algorithms such as the effective QFR set calculation, the hole drilling algorithm and the iterative scaling algorithm are used by reactive approaches, which makes it inefficient. In this paper, we propose a novel cardinality estimation approach by combining proactive approach with QFRs. In our approach, data scan will be executed only once to construct the initial first-level histogram. And then, corresponding to all buckets of the first-level histogram, second-level histograms will be constructed and updated based on QFRs. The existence of second-level histograms and the elaborated mechanism dealing with the data update problem can improve the accuracy of cardinality estimation remarkably. Different from the traditional histogram-based approaches, we do not construct only one big histogram covering the whole value range of an attribute, but construct a serials of small second-level histograms covering different parts of the whole value range. These second-level histograms can be constructed and updated independently over time to ensure the performance of the approach. Extensive comparison experiments fully demonstrate the advantages of our approach in accuracy and performance.

On the complexity of sampling query feedback restricted database repair of functional dependency violations

An inconsistent database is a database instance violating integrity constraints. A repair of an inconsistent database is a maximal consistent subset. Sampling from the repair space is an alternative approach meeting the needs of many applications. In this paper, we introduce a new class of repair, query feedback restricted repair, based on the feedback on user's witness query. We first map out a picture of both data and combined complexities of repair existence problems under different cases to identify the intractable cases. Especially, we show that if the query is a projection or a union query, then the decision problem is NP-complete; even worse, if the query is a conjunctive query, the decision problem becomes Σ2P-complete. However, we prove that the combined complexity of the repair existence problem is in LOGSPACE when the witness query is a selection-join query, and this conclusion also implies that the combined complexity of side-effect free deletion propagation problem under group-deletion is in LOGSPACE which is not considered in previous works. Additionally, we provide a polynomial random repair sampling algorithm under combined complexity. At last, we revisit the key preserving condition [1] and show that it will simplify the problem, i.e., some cases become tractable for certain key preserving views, as opposed to their counterparts that are not key preserving.

An Information Search Algorithm Based on Query Delegation and Feedback Delegation in Delay-Tolerant Networks

In this paper, we propose a new three-hop information search algorithm in delay-tolerant networks. In this algorithm, information search is divided into two phases, query and feedback. In the query...

A Partial Histogram Maintenance Mode

Although histogram has been introduced into the optimizer of the relational database to assist cardinality estimates for more than thirty years, the issue of histogram maintenance is not addressed well. The application of the query feedback information provides a new way of thinking for the histogram maintenance. But it is not mature until now and inefficiency is the main problem. The query feedback-based histogram maintenance approach adopts the entire update mode and the whole histogram will be updated in each maintenance process. This mode brings huge computation for the maintenance process and becomes the most important reason for its inefficiency. Therefore, we propose the partial update mode in this paper to replace the entire update mode and try to improve the maintenance efficiency by limiting the updating range and decreasing the computation in each maintenance process.

Kazakh Concept Query Comparison Based on Pseudo Relevance Feedback Query Expansion Algorithm

Kazakh Concept Query Comparison Based on Pseudo Relevance Feedback Query Expansion Algorithm

Memory-limited non-U-shaped learning with solved open problems

In empirical cognitive science, for human learning, a semantic or behavioral U-shape occurs when a learner first learns, then unlearns, and, finally, relearns, some target concept.Within the formal framework of Inductive Inference, for learning from positive data, previous results have shown, for example, that such U-shapes are unnecessary for explanatory learning, but are necessary for behaviorally correct and non-trivial vacillatory learning.Herein we also distinguish between semantic and syntactic U-shapes. We answer a number of open questions in the prior literature as well as provide new results regarding syntactic U-shapes. Importantly for cognitive science, we see more of a previously noticed pattern that, for parameterized learning criteria, beyond very few initial parameter values, U-shapes are necessary for full learning power.We analyze the necessity of U-shapes in two memory-limited settings.The first setting is Bounded Memory State (BMS) learning, where a learner has an explicitly-bounded state memory, and otherwise only knows its current datum. We show that there are classes learnable with three (or more) memory states that are not learnable non-U-shapedly with any finite number of memory states. This result is surprising, since, for learning with one or two memory states, U-shapes are known to be unnecessary. This solves an open question from the literature.The second setting is that of Memoryless Feedback (MLF) learning, where a learner may ask a bounded number of questions about what data has been seen so far, and otherwise only knows its current datum. We show that there is a class learnable memorylessly with a single feedback query such that this class is not learnable non-U-shapedly memorylessly with any finite number of feedback queries.We employ self-learning classes together with the Operator Recursion Theorem for many of our results, but we also introduce two new techniques for obtaining results. The first is for transferring inclusion results from one setting to another. The main part of the second is the Hybrid Operator Recursion Theorem, which enables us to separate some learning criteria featuring complexity-bounded learners, employing self-learning classes. Both techniques are not specific to U-shaped learning, but applicable for a wide range of settings.

Salton and Buckley’s Landmark Research in Experimental Text Information Retrieval

Salton and Buckley’s Landmark Research in Experimental Text Information Retrieval

Towards a context sensitive approach to searching information based on domain specific knowledge sources

In the context of document retrieval in the biomedical domain, this paper introduces a novel approach to searching for biomedical information using contextual semantic information. More specifically, we propose to combine the contextual semantic information in documents and user queries in an attempt to improve the performance of biomedical information retrieval (IR) systems. Contextual information provides knowledge about a domain in a global context or statistical properties of a sub collection of documents related to a given query in a local context. In our context sensitive IR approach, terms denoting concepts are extracted from each document using several biomedical terminologies. Preferred terms denoting concepts are used to enrich the semantics of the document content via document expansion. The user query is expanded using terms extracted from the top-ranked expanded documents via a blind feedback query expansion approach. In addition, we aim to evaluate the utility of incorporating several terminologies within the proposed context sensitive approach. The experiments carried out on the TREC Genomics 2004 and 2005 test sets show that our context-sensitive IR approach significantly outperforms state-of-the-art baseline approaches.

Towards a Context Sensitive Approach to Searching Information Based on Domain Specific Knowledge Sources

In the context of document retrieval in the biomedical domain, this paper introduces a novel approach to searching for biomedical information using contextual semantic information. More specifically, we propose to combine the contextual semantic information in documents and user queries in an attempt to improve the performance of biomedical information retrieval (IR) systems. Contextual information provides knowledge about a domain in a global context or statistical properties of a sub collection of documents related to a given query in a local context. In our context sensitive IR approach, terms denoting concepts are extracted from each document using several biomedical terminologies. Preferred terms denoting concepts are used to enrich the semantics of the document content via document expansion. The user query is expanded using terms extracted from the top-ranked expanded documents via a blind feedback query expansion approach. In addition, we aim to evaluate the utility of incorporating several terminologies within the proposed context sensitive approach. The experiments carried out on the TREC Genomics 2004 and 2005 test sets show that our context-sensitive IR approach significantly outperforms state-of-the-art baseline approaches.