arXiv subjects

Ke Yi

Publications and source records attributed to Ke Yi.

At least 19 recordsLinked to original sources

arXiv2026

Function Privatization in the Local Model

We study the problem of privately releasing functions, with a particular focus on curves, which are images of continuous functions on some finite interval. Many types of data exist naturally as curves, such as trajectory data or $1$D density curves. We shall primarily be interested in the local model setting, where the function to be privatized captures data belonging to one individual, which is the more challenging setting with limited prior work. Under the standard notion of local differential privacy (DP), any two arbitrarily different functions are required to be made indistinguishable by privatization, which is too strong to allow meaningful utility; we thus work with a generalized notion of DP known as Geo-Privacy (GP), which allows functions far apart to be distinguished more easily while providing strong protection for near functions. To demonstrate the effectiveness of our framework, we provide experimental evaluation on several datasets.

cs.CR↗

arXiv2026

Adaptive Privacy Budgeting

We study the problem of adaptive privacy budgeting under generalized differential privacy. Consider the setting where each user $i\in [n]$ holds a tuple $x_i\in U:=U_1\times \dotsb \times U_T$, where $x_i(l)\in U_l$ represents the $l$-th component of their data. For every $l\in [T]$ (or a subset), an untrusted analyst wishes to compute some $f_l(x_1(l),\dots,x_n(l))$, while respecting the privacy of each user. For many functions $f_l$, data from the users are not all equally important, and there is potential to use the privacy budgets of the users strategically, leading to privacy savings that can be used to improve the utility of later queries. In particular, the budgeting should be adaptive to the outputs of previous queries, so that greater savings can be achieved on more typical instances. In this paper, we provide such an adaptive budgeting framework, with various applications demonstrating its applicability.

cs.CR↗

arXiv2025

Approximate DBSCAN under Differential Privacy

This paper revisits the DBSCAN problem under differential privacy (DP). Existing DP-DBSCAN algorithms aim at publishing the cluster labels of the input points. However, we show that both empirically and theoretically, this approach cannot offer any utility in the published results. We therefore propose an alternative definition of DP-DBSCAN based on the notion of spans. We argue that publishing the spans actually better serves the purposes of visualization and classification of DBSCAN. Then we present a linear-time DP-DBSCAN algorithm achieving the sandwich quality guarantee in any constant dimensions, as well as matching lower bounds on the approximation ratio. A key building block in our algorithm is a linear-time algorithm for constructing a histogram under pure-DP, which is of independent interest. Finally, we conducted experiments on both synthetic and real-world datasets to verify the practical performance of our DP-DBSCAN algorithm.

cs.CR↗

arXiv2025

Yannakakis+: Practical Acyclic Query Evaluation with Theoretical Guarantees

Acyclic conjunctive queries form the backbone of most analytical workloads, and have been extensively studied in the literature from both theoretical and practical angles. However, there is still a large divide between theory and practice. While the 40-year-old Yannakakis algorithm has strong theoretical running time guarantees, it has not been adopted in real systems due to its high hidden constant factor. In this paper, we strive to close this gap by proposing Yannakakis+, an improved version of the Yannakakis algorithm, which is more practically efficient while preserving its theoretical guarantees. Our experiments demonstrate that Yannakakis+ consistently outperforms the original Yannakakis algorithm by 2x to 5x across a wide range of queries and datasets. Another nice feature of our new algorithm is that it generates a traditional DAG query plan consisting of standard relational operators, allowing Yannakakis+ to be easily plugged into any standard SQL engine. Our system prototype currently supports four different SQL engines (DuckDB, PostgreSQL, SparkSQL, and AnalyticDB from Alibaba Cloud), and our experiments show that Yannakakis+ is able to deliver better performance than their native query plans on 160 out of the 162 queries tested, with an average speedup of 2.41x and a maximum speedup of 47,059x.

cs.DB↗

arXiv2024

Rotated Runtime Smooth: Training-Free Activation Smoother for accurate INT4 inference

Large language models have demonstrated promising capabilities upon scaling up parameters. However, serving large language models incurs substantial computation and memory movement costs due to their large scale. Quantization methods have been employed to reduce service costs and latency. Nevertheless, outliers in activations hinder the development of INT4 weight-activation quantization. Existing approaches separate outliers and normal values into two matrices or migrate outliers from activations to weights, suffering from high latency or accuracy degradation. Based on observing activations from large language models, outliers can be classified into channel-wise and spike outliers. In this work, we propose Rotated Runtime Smooth (RRS), a plug-and-play activation smoother for quantization, consisting of Runtime Smooth and the Rotation operation. Runtime Smooth (RS) is introduced to eliminate channel-wise outliers by smoothing activations with channel-wise maximums during runtime. The rotation operation can narrow the gap between spike outliers and normal values, alleviating the effect of victims caused by channel-wise smoothing. The proposed method outperforms the state-of-the-art method in the LLaMA and Qwen families and improves WikiText-2 perplexity from 57.33 to 6.66 for INT4 inference.

cs.LG↗

arXiv2024

One QuantLLM for ALL: Fine-tuning Quantized LLMs Once for Efficient Deployments

Large Language Models (LLMs) have advanced rapidly but face significant memory demands. While quantization has shown promise for LLMs, current methods typically require lengthy training to alleviate the performance degradation from quantization loss. However, deploying LLMs across diverse scenarios with different resource constraints, e.g., servers and personal computers, requires repeated training per application, which amplifies the lengthy training problem. Given that, it is advantageous to train a once-for-all (OFA) supernet capable of yielding diverse optimal subnets for downstream applications through one-shot training. Nonetheless, the scale of current language models impedes efficiency and amplifies interference from weight sharing between subnets. We make an initial attempt to extend the once-for-all framework to large language models. Specifically, we decouple shared weights to eliminate the interference and incorporate Low-Rank adapters for training efficiency. Furthermore, we observe the imbalance allocation of training resources from the traditional uniform sampling. A non-parametric scheduler is introduced to adjust the sampling rate for each quantization configuration, achieving a more balanced allocation among subnets with varying demands. We validate the approach on LLaMA2 families, and downstream evaluation confirms our ability to maintain high performance while significantly reducing deployment time faced with multiple scenarios.

cs.AI↗

arXiv2024

Smooth Sensitivity for Geo-Privacy

Suppose each user $i$ holds a private value $x_i$ in some metric space $(U, \mathrm{dist})$, and an untrusted data analyst wishes to compute $\sum_i f(x_i)$ for some function $f : U \rightarrow \mathbb{R}$ by asking each user to send in a privatized $f(x_i)$. This is a fundamental problem in privacy-preserving population analytics, and the local model of differential privacy (LDP) is the predominant model under which the problem has been studied. However, LDP requires any two different $x_i, x'_i$ to be $\varepsilon$-distinguishable, which can be overly strong for geometric/numerical data. On the other hand, Geo-Privacy (GP) stipulates that the level of distinguishability be proportional to $\mathrm{dist}(x_i, x_i')$, providing an attractive alternative notion of personal data privacy in a metric space. However, existing GP mechanisms for this problem, which add a uniform noise to either $x_i$ or $f(x_i)$, are not satisfactory. In this paper, we generalize the smooth sensitivity framework from Differential Privacy to Geo-Privacy, which allows us to add noise tailored to the hardness of the given instance. We provide definitions, mechanisms, and a generic procedure for computing the smooth sensitivity under GP equipped with a general metric. Then we present three applications: one-way and two-way threshold functions, and Gaussian kernel density estimation, to demonstrate the applicability and utility of our smooth sensitivity framework.

cs.CR↗

arXiv2024

Reservoir Sampling over Joins

Sampling over joins is a fundamental task in large-scale data analytics. Instead of computing the full join results, which could be massive, a uniform sample of the join results would suffice for many purposes, such as answering analytical queries or training machine learning models. In this paper, we study the problem of how to maintain a random sample over joins while the tuples are streaming in. Without the join, this problem can be solved by some simple and classical reservoir sampling algorithms. However, the join operator makes the problem significantly harder, as the join size can be polynomially larger than the input. We present a new algorithm for this problem that achieves a near-linear complexity. The key technical components are a generalized reservoir sampling algorithm that supports a predicate, and a dynamic index for sampling over joins. We also conduct extensive experiments on both graph and relational data over various join queries, and the experimental results demonstrate significant performance improvement over the state of the art.

cs.DB↗

arXiv2024

Secure Query Processing with Linear Complexity

We present LINQ, the first join protocol with linear complexity (in both running time and communication) under the secure multi-party computation model (MPC). It can also be extended to support all free-connex queries, a large class of select-join-aggregate queries, still with linear complexity. This matches the plaintext result for the query processing problem, as free-connex queries are the largest class of queries known to be solvable in linear time in plaintext. We have then built a query processing system based on LINQ, and the experimental results show that LINQ significantly outperforms the state of the art. For example, it can finish a query on three relations with an output size of 1 million tuples in around 100s in the LAN setting, while existing protocols that support the query cannot finish in an hour. Thus LINQ brings MPC query processing closer to practicality.

cs.CR↗

arXiv2024

Almost Instance-optimal Clipping for Summation Problems in the Shuffle Model of Differential Privacy

Differentially private mechanisms achieving worst-case optimal error bounds (e.g., the classical Laplace mechanism) are well-studied in the literature. However, when typical data are far from the worst case, \emph{instance-specific} error bounds -- which depend on the largest value in the dataset -- are more meaningful. For example, consider the sum estimation problem, where each user has an integer $x_i$ from the domain $\{0,1,\dots,U\}$ and we wish to estimate $\sum_i x_i$. This has a worst-case optimal error of $O(U/\varepsilon)$, while recent work has shown that the clipping mechanism can achieve an instance-optimal error of $O(\max_i x_i \cdot \log\log U /\varepsilon)$. Under the shuffle model, known instance-optimal protocols are less communication-efficient. The clipping mechanism also works in the shuffle model, but requires two rounds: Round one finds the clipping threshold, and round two does the clipping and computes the noisy sum of the clipped data. In this paper, we show how these two seemingly sequential steps can be done simultaneously in one round using just $1+o(1)$ messages per user, while maintaining the instance-optimal error bound. We also extend our technique to the high-dimensional sum estimation problem and sparse vector aggregation (a.k.a. frequency estimation under user-level differential privacy).

cs.CR↗

arXiv2023

Concentrated Geo-Privacy

This paper proposes concentrated geo-privacy (CGP), a privacy notion that can be considered as the counterpart of concentrated differential privacy (CDP) for geometric data. Compared with the previous notion of geo-privacy [ABCP13, CABP13], which is the counterpart of standard differential privacy, CGP offers many benefits including simplicity of the mechanism, lower noise scale in high dimensions, and better composability known as advanced composition. The last one is the most important, as it allows us to design complex mechanisms using smaller building blocks while achieving better utilities. To complement this result, we show that the previous notion of geo-privacy inherently does not admit advanced composition even using its approximate version. Next, we study three problems on private geometric data: the identity query, k nearest neighbors, and convex hulls. While the first problem has been previously studied, we give the first mechanisms for the latter two under geo-privacy. For all three problems, composability is essential in obtaining good utility guarantees on the privatized query answer.

cs.CR↗

arXiv2023

Change Propagation Without Joins

We revisit the classical change propagation framework for query evaluation under updates. The standard framework takes a query plan and materializes the intermediate views, which incurs high polynomial costs in both space and time, with the join operator being the culprit. In this paper, we propose a new change propagation framework without joins, thus naturally avoiding this polynomial blowup. Meanwhile, we show that the new framework still supports constant-delay enumeration of both the deltas and the full query results, the same as in the standard framework. Furthermore, we provide a quantitative analysis of its update cost, which not only recovers many recent theoretical results on the problem, but also yields an effective approach to optimizing the query plan. The new framework is also easy to be integrated into an existing streaming database system. Experimental results show that our system prototype, implemented using Flink DataStream API, significantly outperforms other systems in terms of space, time, and latency.

cs.DB↗

arXiv2022

Differential Privacy on Dynamic Data

A fundamental problem in differential privacy is to release a privatized data structure over a dataset that can be used to answer a class of linear queries with small errors. This problem has been well studied in the static case. In this paper, we consider the dynamic setting where items may be inserted into or deleted from the dataset over time, and we need to continually release data structures so that queries can be answered at any time. We present black-box constructions of such dynamic differentially private mechanisms from static ones with only a polylogarithmic degradation in the utility. For the fully-dynamic case, this is the first such result. For the insertion-only case, similar constructions are known, but we improve them over sparse update streams.

cs.CR↗

arXiv2022

Differentially Private Covariance Revisited

In this paper, we present two new algorithms for covariance estimation under concentrated differential privacy (zCDP). The first algorithm achieves a Frobenius error of $\tilde{O}(d^{1/4}\sqrt{\mathrm{tr}}/\sqrt{n} + \sqrt{d}/n)$, where $\mathrm{tr}$ is the trace of the covariance matrix. By taking $\mathrm{tr}=1$, this also implies a worst-case error bound of $\tilde{O}(d^{1/4}/\sqrt{n})$, which improves the standard Gaussian mechanism's $\tilde{O}(d/n)$ for the regime $d>\widetilde{\Omega}(n^{2/3})$. Our second algorithm offers a tail-sensitive bound that could be much better on skewed data. The corresponding algorithms are also simple and efficient. Experimental results show that they offer significant improvements over prior work.

cs.CR↗

arXiv2022

OneDConv: Generalized Convolution For Transform-Invariant Representation

Convolutional Neural Networks (CNNs) have exhibited their great power in a variety of vision tasks. However, the lack of transform-invariant property limits their further applications in complicated real-world scenarios. In this work, we proposed a novel generalized one dimension convolutional operator (OneDConv), which dynamically transforms the convolution kernels based on the input features in a computationally and parametrically efficient manner. The proposed operator can extract the transform-invariant features naturally. It improves the robustness and generalization of convolution without sacrificing the performance on common images. The proposed OneDConv operator can substitute the vanilla convolution, thus it can be incorporated into current popular convolutional architectures and trained end-to-end readily. On several popular benchmarks, OneDConv outperforms the original convolution operation and other proposed models both in canonical and distorted images.

cs.CV↗

arXiv2021

A Nearly Instance-optimal Differentially Private Mechanism for Conjunctive Queries

Releasing the result size of conjunctive queries and graph pattern queries under differential privacy (DP) has received considerable attention in the literature, but existing solutions do not offer any optimality guarantees. We provide the first DP mechanism for this problem with a fairly strong notion of optimality, which can be considered as a natural relaxation of instance-optimality to a constant.

cs.DB↗

arXiv2021

Frequency Estimation in the Shuffle Model with Almost a Single Message

We present a protocol in the shuffle model of differential privacy (DP) for the \textit{frequency estimation} problem that achieves error $\omega(1)\cdot O(\log n)$, almost matching the central-DP accuracy, with $1+o(1)$ messages per user. This exhibits a sharp transition phenomenon, as there is a lower bound of $\Omega(n^{1/4})$ if each user is allowed to send only one message. Previously, such a result is only known when the domain size $B$ is $o(n)$. For a large domain, we also need an efficient method to identify the \textit{heavy hitters} (i.e., elements that are frequent enough). For this purpose, we design a shuffle-DP protocol that uses $o(1)$ messages per user and can identify all heavy hitters in time polylogarithmic in $B$. Finally, by combining our frequency estimation and the heavy hitter detection protocols, we show how to solve the $B$-dimensional \textit{1-sparse vector summation} problem in the high-dimensional setting $B=\Omega(n)$, achieving the optimal central-DP MSE $\tilde O(n)$ with $1+o(1)$ messages per user. In addition to error and message number, our protocols improve in terms of message size and running time as well. They are also very easy to implement. The experimental results demonstrate order-of-magnitude improvement over prior work.

cs.CR↗

arXiv2021

Universal Private Estimators

We present \textit{universal} estimators for the statistical mean, variance, and scale (in particular, the interquartile range) under pure differential privacy. These estimators are universal in the sense that they work on an arbitrary, unknown continuous distribution $\mathcal{P}$ over $\mathbb{R}$, while yielding strong utility guarantees except for ill-behaved $\mathcal{P}$. For certain distribution families like Gaussians or heavy-tailed distributions, we show that our universal estimators match or improve existing estimators, which are often specifically designed for the given family and under \textit{a priori} boundedness assumptions on the mean and variance of $\mathcal{P}$. This is the first time these boundedness assumptions are removed under pure differential privacy. The main technical tools in our development are instance-optimal empirical estimators for the mean and quantiles over the unbounded integer domain, which can be of independent interest.

cs.CR↗