arXiv subjects

Chengcheng Yang

Publications and source records attributed to Chengcheng Yang.

14 recordsLinked to original sources

arXiv2026

Setoka: A Benchmark for Hierarchical User Understanding in Personalized Agents over Heterogeneous Data

Personalized agents are increasingly applied to assist users across a wide range of tasks. Effective personalized assistance requires not only retrieving explicit facts from past interactions stored in agent memory, but also inferring abstract personal characteristics. However, existing memory benchmarks primarily evaluate whether an agent can retrieve information explicitly stated in conversational histories, failing to provide an effective assessment of deeper user understanding. In this work, we propose Setoka, a benchmark for evaluating memory-augmented personalized agents with hierarchical user understanding from heterogeneous data. Grounded in theories from cognitive and personality psychology, Setoka defines four levels of user understanding, i.e., semantic memory, episodic memory, behavior pattern, and personality trait. Moreover, to enable realistic yet privacy-preserving evaluation, we design a psychometrics-based pipeline that synthesizes diverse, coherent heterogeneous user data and queries at scale. Finally, we leverage Setoka to evaluate 3 language models combined with 5 memory systems for 10 synthetic users. Our comprehensive evaluation reveals that while existing systems perform well on semantic memory retrieval, their performance declines on episodic memory. Moreover, when dealing with behavior pattern and personality trait understanding tasks that require integrating heterogeneous and fragmented information dispersed over time, performance declines even further. These findings demonstrate that user understanding cannot be handled by simple fact retrieval, motivating the design of memory mechanisms for cross-source integration and abstraction over long-term user behavior.

cs.AI↗

arXiv2025

MOSAIC: Codon Harmonization of Monte Carlo-Based Simulated Annealing for Linked Codons in Heterologous Protein Expression

Codon usage bias has a crucial impact on the translation efficiency and co-translational folding of proteins, necessitating the algorithmic development of codon optimization/harmonization methods, particularly for heterologous recombinant protein expression. Codon harmonization is especially valuable for proteins sensitive to translation rates, because it can potentially replicate native translation speeds, preserving proper folding and maintaining protein activity. This work proposes a Monte Carlo-based codon harmonization algorithm, MOSAIC (Monte Carlo-based Simulated Annealing for Linked Codons), for the harmonization of a set of linked codons, which differs from conventional codon harmonization, by focusing on the codon sets rather than individual ones. Our MOSAIC demonstrates robust computational performance on ribosomal proteins (S18, S15, S10, and L11) as model systems. Among them, the harmonized gene of RP S18 was expressed and compared with the expression of the wild-type gene. The harmonized gene clearly yielded a larger quantity of the protein, from which the amount of the soluble protein was also significant. These results underscored the potential of the linked codon harmonization approach to enhance the expression and functionality of sensitive proteins, setting the stage for more efficient production of recombinant proteins in various biotechnological and pharmaceutical applications.

q-bio.QM↗

arXiv2025

AQETuner: Reliable Query-level Configuration Tuning for Analytical Query Engines

Modern analytical query engines (AQEs) are essential for large-scale data analysis and processing. These systems usually provide numerous query-level tunable knobs that significantly affect individual query performance. While several studies have explored automatic DBMS configuration tuning, they have several limitations to handle query-level tuning. Firstly, they fail to capture how knobs influence query plans, which directly affect query performance. Secondly, they overlook query failures during the tuning processing, resulting in low tuning efficiency. Thirdly, they struggle with cold-start problems for new queries, leading to prolonged tuning time. To address these challenges, we propose AQETuner, a novel Bayesian Optimization-based system tailored for reliable query-level knob tuning in AQEs. AQETuner first applies the attention mechanisms to jointly encode the knobs and plan query, effectively identifying the impact of knobs on plan nodes. Then, AQETuner employs a dual-task Neural Process to predict both query performance and failures, leveraging their interactions to guide the tuning process. Furthermore, AQETuner utilizes Particle Swarm Optimization to efficiently generate high-quality samples in parallel during the initial tuning stage for the new queries. Experimental results show that AQETuner significantly outperforms existing methods, reducing query latency by up to 23.7% and query failures by up to 51.2%.

cs.DB↗

arXiv2025

Learning from the Past: Adaptive Parallelism Tuning for Stream Processing Systems

Distributed stream processing systems rely on the dataflow model to define and execute streaming jobs, organizing computations as Directed Acyclic Graphs (DAGs) of operators. Adjusting the parallelism of these operators is crucial to handling fluctuating workloads efficiently while balancing resource usage and processing performance. However, existing methods often fail to effectively utilize execution histories or fully exploit DAG structures, limiting their ability to identity bottlenecks and determine the optimal parallelism. In this paper, we propose StreamTune, a novel approach for adaptive paralelism tuning in stream processing systems. StreamTune incorporates a pre-training and fine-tuning framework that leverages global knowledge from historical execution data for job-specific parallelism tuning. In the pre-training phase, Stream Tune clusters the historical data with Graph Edit Distance and pre-trains a Graph Neural Networkbased encoder per cluster to capture the correlation between the operator parallelism, DAG structures, and the identified operator-level bottlenecks. In the online tuning phase, StreamTune iteratively refines operator parallelism recommendations using an operator-level bottleneck prediction model enforced with a monotonic constraint, which aligns with the observed system performance behavior. Evaluation results demonstrate that StreamTune reduces reconfigurations by up to 29.6% and parallelism degrees by up to 30.8% in Apache Flink under a synthetic workload. In Timely Dataflow, StreamTune achieves up to an 83.3% reduction in parallelism degrees while maintaining comparable processing performance under the Nexmark benchmark, when compared to the state-of-the-art methods.

cs.DC↗

arXiv2025

The Plateau Problem of Michell Trusses and Orthogonality in Springs

Given finitely many pointed forces in the plane. Suppose that these forces sum up to zero and their net torques also sum up to zero. One can show that there exists a system of springs whose boundary forces exactly counter-balance these pointed forces. We will generalize to higher dimensions using the Cauchy stress tensor for elastic materials. Given a system of springs, we can multiply the length of each spring with its corresponding spring constant and then sum these products up. The result is called the total mass of the system. We are interested in the Plateau problem of the existence of the minimal spring system given a boundary condition. This minimization problem was first introduced in 1904 by A. Michell. He showed that a minimizer could smear out. The Michell Truss became known in mechanical engineering. It raised attention in optimal design, such as minimizing costs in building bridges. In 1960s and 1970s, the problem was developed using PDE and convex analysis by introducing an equivalent dual maximization problem. In 2008, Bouchitt\'{e}, Gangbo, and Sppecher introduced lines of principal actions to generalize Hencky-Prandtle net to higher dimensional duality and proved that the minimizer can be found provided that it exists. In the unpublished notes of Gangbo, he also showed that if springs of the same kind are optimal. In this paper, we are going to solve the Plateau problem using two different tools in GMT: first, a minimizer can be viewed as a flat chain complex; second, a minimizer can also be viewed as a current. At the end, we are going to show one progress in discovering the topological properties of minimizers: compressed and stretched springs must be perpendicular to each other at non-boundary points. I appreciate my advisor Prof. Robert Hardt for communicating with me regularly on this problem.

math.OC↗

arXiv2024

EasyTime: Time Series Forecasting Made Easy

Time series forecasting has important applications across diverse domains. EasyTime, the system we demonstrate, facilitates easy use of time-series forecasting methods by researchers and practitioners alike. First, EasyTime enables one-click evaluation, enabling researchers to evaluate new forecasting methods using the suite of diverse time series datasets collected in the preexisting time series forecasting benchmark (TFB). This is achieved by leveraging TFB's flexible and consistent evaluation pipeline. Second, when practitioners must perform forecasting on a new dataset, a nontrivial first step is often to find an appropriate forecasting method. EasyTime provides an Automated Ensemble module that combines the promising forecasting methods to yield superior forecasting accuracy compared to individual methods. Third, EasyTime offers a natural language Q&A module leveraging large language models. Given a question like "Which method is best for long term forecasting on time series with strong seasonality?", EasyTime converts the question into SQL queries on the database of results obtained by TFB and then returns an answer in natural language and charts. By demonstrating EasyTime, we intend to show how it is possible to simplify the use of time series forecasting and to offer better support for the development of new generations of time series forecasting methods.

cs.LG↗

arXiv2024

ByteCard: Enhancing ByteDance's Data Warehouse with Learned Cardinality Estimation

Cardinality estimation is a critical component and a longstanding challenge in modern data warehouses. ByteHouse, ByteDance's cloud-native engine for extensive data analysis in exabyte-scale environments, serves numerous internal decision-making business scenarios. With the increasing demand for ByteHouse, cardinality estimation becomes the bottleneck for efficiently processing queries. Specifically, the existing query optimizer of ByteHouse uses the traditional Selinger-like cardinality estimator, which can produce substantial estimation errors, resulting in suboptimal query plans. To improve cardinality estimation accuracy while maintaining a practical inference overhead, we develop a framework ByteCard that enables efficient training and integration of learned cardinality estimators. Furthermore, ByteCard adapts recent advances in cardinality estimation to build models that can balance accuracy and practicality (e.g., inference latency, model size, training overhead). We observe significant query processing speed-up in ByteHouse after replacing the existing cardinality estimator with ByteCard for several optimization scenarios. Evaluations on real-world datasets show the integration of ByteCard leads to an improvement of up to 30% in the 99th quantile of latency. At last, we share our valuable experience in engineering advanced cardinality estimators. This experience can help ByteHouse integrate more learning-based solutions on the critical query execution path in the future.

cs.DB↗

arXiv2024

Michell Truss and From 1-beam to k-beam

This paper generalizes the Michell Truss problem and Gangbo's paper from 1-dimension to higher dimensions using geometric measure theory. Given an elastic surface $S$ made of $(k-1)$-beams under an equilibriated system $F$ of external forces, then we ask the following two questions: 1. What are the necessary and sufficient conditions for the existence of an elastic body made of $k$-beams whose forces on the surface balance $F$ and whose surfaces consist of $S$. 2. What is an optimal design so that the total cost is a minimum? We've solved the existence question completely; and research is still in progress for the minimal question. In particular when $k=1$, it involves a system of beams joining a given finite collection of pointed forces. It was first introduced by A. Michell in 1904, then used in mechanical engineering, and recently popularized in many pure mathematics works by W. Gangbo, Prager, and others. Here we are going to generalize them to higher dimensional cases. We have already found the minimal solutions in terms of the flat chain complex and vector-valued currents. Right now we are studying the Calibration theory for future directions. I appreciate the discussion with Prof. Robert Hardt!

math.OC↗

arXiv2023

Geodesics in 3-dimensional Euclidean Space with One or Two Analytic Obstacles

In many singular metric spaces, the regularity of a shortest-length curve is unknown. Algebraic varieties, or more generally sets defined by finitely many polynomial or real analytic equalities or inequalities, all locally partition into finitely many real analytic submanifolds called strata. So any component of a shortest-length curve which lies completely in one such stratum is a geodesic in the stratum, hence an embedded real analytic curve. The key question thus is whether there are only finitely many components. F. Albrecht and I.D. Berg proved this is true for a geodesic in a closed region of $n$-dimensional Euclidean space with a smooth real analytic hyper surface as boundary. Here the curve consists of finitely many interior line segments alternating with boundary hypersurface geodesics. Their bound on the number of these depended on the initial velocity of the geodesic, and they conjectured that is independent. Here we prove this independence in $\mathbb{R}^3$. We also generalize their result to regions whose boundary is locally the boundary of the union of two transversally intersecting analytic hypersurfaces in $\mathbb{R}^3$.

math.DG↗

arXiv2021

A Geodesic Stratification of Two-dimensional Semi-algebraic Sets

Given any arbitrary semi-algebraic set $X$, any two points in $X$ may be joined by a piecewise $C^2$ path $γ$ of shortest length. Suppose $\mathcal{A}$ is a semi-algebraic stratification of $X$ such that each component of $γ\cap \mathcal{A}$ is either a singleton or a real analytic geodesic segment in $\mathcal{A}$, the question is whether $γ\cap \mathcal{A}$ has at most finitely many such components. This paper gives a semi-algebraic stratification, in particular a cell decomposition, of a real semi-algebraic set in the plane whose open cells have this finiteness property. This provides insights for high dimensional stratifications of semi-algebraic sets in connection with geodesics.

math.AG↗

arXiv2020

A triangulation of semi-algebraic sets concerning an analytical condition for shortest-length curves

This paper concerns an analytical stratification question of real algebraic and semi-algebraic sets. For Whitney's stratification in 1957, it partitions a real algebraic set into partial algebraic manifolds\cite{W}. In 1975 Hironaka reproved that a real algebraic set is triangulable and also generalized it to semi-algebraic sets, following the idea of Lojasiewicz's triangulation of semi-analytic sets in 1964. Following their examples and wondering how geometry looks like locally. this paper tries to come up with a stratification, in particular a cell decomposition, such that it satisfies the following analytical property. Given any shortest curve between two points in a real algebraic or semi-algebraic set, it interacts each cell (or simplex) at most finitely many times.

math.AG↗

arXiv2020

A Problem of Erdös Concerning Lattice Cubes

This paper studies a problem of Erdös concerning lattice cubes. Given an $N \times N \times N$ lattice cube, we want to find the maximum number of vertices one can select so that no eight corners of a rectangular box are chosen simultaneously. Erdös conjectured that it has a sharp upper bound, which is $O(N^{11/4})$, but no example that large has been found yet. We start approaching this question for small $N$ using the method of exhaustion, and we find that there is not necessarily a unique maximal set of vertices (counting all possible symmetries). Next, we study an equivalent two-dimensional version of this problem looking for patterns that might be useful for generalizing to the three-dimensional case. Since an $n \times n$ grid is also an $n \times n$ matrix, we rephrase and generalize the original question to: what is the minimum number $α(k,n)$ of vertices one can put in an $n \times n$ matrix with entries 0 and 1, such that every $k \times k$ minor contains at least one entry of 1, for $1 \leq k \leq n$? We discover some interesting formulas and asymptotic patterns that shed new light on the question.

math.CO↗

arXiv2014

Gravity Effects on Information Filtering and Network Evolving

In this paper, based on the gravity principle of classical physics, we propose a tunable gravity-based model, which considers tag usage pattern to weigh both the mass and distance of network nodes. We then apply this model in solving the problems of information filtering and network evolving. Experimental results on two real-world data sets, \emph{Del.icio.us} and \emph{MovieLens}, show that it can not only enhance the algorithmic performance, but can also better characterize the properties of real networks. This work may shed some light on the in-depth understanding of the effect of gravity model.

physics.soc-ph↗

arXiv2012

A two-step Recommendation Algorithm via Iterative Local Least Squares

Recommender systems can change our life a lot and help us select suitable and favorite items much more conveniently and easily. As a consequence, various kinds of algorithms have been proposed in last few years to improve the performance. However, all of them face one critical problem: data sparsity. In this paper, we proposed a two-step recommendation algorithm via iterative local least squares (ILLS). Firstly, we obtain the ratings matrix which is constructed via users' behavioral records, and it is normally very sparse. Secondly, we preprocess the "ratings" matrix through ProbS which can convert the sparse data to a dense one. Then we use ILLS to estimate those missing values. Finally, the recommendation list is generated. Experimental results on the three datasets: MovieLens, Netflix, RYM, suggest that the proposed method can enhance the algorithmic accuracy of AUC. Especially, it performs much better in dense datasets. Furthermore, since this methods can improve those missing value more accurately via iteration which might show light in discovering those inactive users' purchasing intention and eventually solving cold-start problem.

cs.IR↗