About

Lavi Peng is an Assistant Professor in the Hospitality & Tourism Management department at the Isenberg School of Management, University of Massachusetts Amherst. He earned his PhD in Hospitality Management from Pennsylvania State University in 2025, along with a Master of Science in International Hospitality Management from Hong Kong Polytechnic University in 2019, and a Bachelor of Arts in International Hotel Management from Vatel Hotel & Tourism Business School in France in 2017. His research interests focus on consumer well-being, foodservice management, and service technology. Peng has contributed to the field through various publications exploring topics such as gendered robots and persuasion, consumer responses to sociopolitical activism in hospitality, and the influence of environmental cues on consumer behavior. His work emphasizes understanding consumer behavior and service experiences within the hospitality industry.

Research topics

• Computer Science
• Artificial Intelligence
• Machine Learning
• Data Mining
• Geography
• Business
• Risk analysis (engineering)

Selected publications

• Multi-entity co-simulation of intelligent vehicle based on distributed message-oriented middleware

  Journal of Highway and Transportation Research and Development (English Edition) · 2025-03-01

  articleOpen access1st authorCorresponding

  With the integration of cutting-edge technologies such as information communication, the internet, big data, cloud computing, and artificial intelligence into intelligent vehicles, the scale and complexity of their cyber-physical subsystems have been steadily increasing. In response to the growing demands for high concurrent access and efficient operation within the multiple entities of information system modeling and simulation, this study introduces a novel method for multi-entity co-simulation of intelligent vehicles, grounded in distributed message-oriented middleware. Tailored for the unique challenges of multi-entity co-simulation scenarios, the proposed method involves classifying heterogeneous simulation platforms and developing distinct implementation modes for message middleware interface units corresponding to each platform type. This approach significantly enhances the efficiency of multi-entity co-simulation in intelligent vehicle systems, facilitating more effective integration and operation of diverse subsystems.

  PublisherOA PDFDOI

• Efficient and robust active learning methods for interactive database exploration

  The VLDB Journal · 2023 · 2 citations

  • Computer Science
  • Computer Science
  • Machine Learning
  PublisherOA PDFDOI

• Supporting Scientific Analytics under Data Uncertainty and Query Uncertainty

  Scholarworks (University of Massachusetts Amherst) · 2021-04-01

  articleOpen access1st authorCorresponding

  Despite a considerable progress in verification and abstraction of random and control logic, advances in formal verification of arithmetic designs have been lagging. This can be attributed mostly to the difficulty in an efficient modeling of arithmetic circuits and datapaths without resorting to computationally expensive Boolean methods, such as Binary Decision Diagrams (BDDs) and Boolean Satisfiability (SAT), that require “bit blasting”, i.e., flattening the design to a bit-level netlist. Approaches that rely on computer algebra and Satisfiability Modulo Theories (SMT) methods are either too abstract to handle the bit-level nature of arithmetic designs or require solving computationally expensive decision or satisfiability problems. The work proposed in this thesis aims at overcoming the limitations of analyzing arithmetic circuits, specifically at the post-synthesized phase. It addresses the verification, abstraction and reverse engineering problems of arithmetic circuits at an algebraic level, treating an arithmetic circuit and its specification as a properly constructed algebraic system. The proposed technique solves these problems by function extraction, i.e., by deriving arithmetic function computed by the circuit from its low-level circuit implementation using computer algebraic rewriting technique. The proposed techniques work on large integer arithmetic circuits and finite field arithmetic circuits, up to 512-bit wide containing millions of logic gates.

  PublisherDOI

• Runway - Model Lifecycle Management at Netflix

  2020 · 2 citations

  Senior authorCorresponding
  • Computer Science
  • Business
  • Computer Science
  Publisher

• Optimization for active learning-based interactive database exploration

  Proceedings of the VLDB Endowment · 2018-09-01 · 21 citations

  articleOpen access

  There is an increasing gap between fast growth of data and limited human ability to comprehend data. Consequently, there has been a growing demand of data management tools that can bridge this gap and help the user retrieve high-value content from data more effectively. In this work, we aim to build interactive data exploration as a new database service, using an approach called "explore-by-example". In particular, we cast the explore-by-example problem in a principled "active learning" framework, and bring the properties of important classes of database queries to bear on the design of new algorithms and optimizations for active learning-based database exploration. These new techniques allow the database system to overcome a fundamental limitation of traditional active learning, i.e., the slow convergence problem. Evaluation results using real-world datasets and user interest patterns show that our new system significantly outperforms state-of-the-art active learning techniques and data exploration systems in accuracy while achieving desired efficiency for interactive performance.

  PublisherOA PDFDOI

• A novel FMEA tool application in semiconductor manufacture

  2017-03-01 · 3 citations

  article

  In this Paper, we introduce a novel FMEA (Failure Mode and effect Analysis) system, It can achieve FMEA more practicable and valuable compared with current FMEA application status as record archives. The novel FMEA basic unit is module and related modules are combined to form a whole FMEA, Six new link/experience function modules are introduced into the standard FMEA format to integrate the database, and the module unit exists independently so that different FMEA file cross share the similar failure modules. Three new link function modules can connect FMEA system with other related production systems to embed FMEA useful resource into production as guidance, this link function can prevent potential and old failure modes timely occurring timely, further reduce defect and improve production efficiency and quality. This novel FMEA tool can make FMEA more value and important functions in semiconductor process.

  PublisherDOI

• Interactive Data Exploration via Machine Learning Models.

  IEEE Data(base) Engineering Bulletin · 2016-01-01 · 21 citations

  articleSenior author
  Publisher

• AIDE

  Proceedings of the VLDB Endowment · 2015-08-01 · 7 citations

  articleSenior author

  Data analysts often engage in data exploration tasks to discover interesting data patterns, without knowing exactly what they are looking for. Such exploration tasks can be very labor-intensive because they often require the user to review many results of ad-hoc queries and adjust the predicates of subsequent queries to balance the tradeoff between collecting all interesting information and reducing the size of returned data. In this demonstration we introduce AIDE , a system that automates these exploration tasks. AIDE steers the user towards interesting data areas based on her relevance feedback on database samples, aiming to achieve the goal of identifying all database objects that match the user interest with high efficiency. In our demonstration, conference attendees will see AIDE in action for a variety of exploration tasks on real-world datasets.

  PublisherDOI

• Supporting Data Uncertainty in Array Databases

  2015-05-27 · 11 citations

  article1st authorCorresponding

  Uncertain data management has become crucial to scientific applications. Recently, array databases have gained popularity for scientific data processing due to performance benefits. In this paper, we address uncertain data management in array databases, which may involve both value uncertainty within individual tuples and position uncertainty regarding where a tuple should belong in an array given uncertain dimension attributes. Our work defines the formal semantics of array operations under both value and position uncertainty. To address the new challenge raised by position uncertainty, we propose a suite of storage and evaluation strategies for array operations, with a focus on a new scheme that bounds the overhead of querying by strategically treating tuples with large variances via replication in storage. Results from real datasets show that for common workloads, our best-performing techniques outperform alternative methods based on state-of-the-art indexes by 1.7x to 4.3x for the Subarray operation and 1 to 2 orders of magnitude for Structure-Join, at only a small storage cost.

  PublisherDOI

• Groupwise analytics via adaptive MapReduce

  2015-04-01 · 1 citations

  article1st authorCorresponding

  Shared-nothing systems such as Hadoop vastly simplify parallel programming when processing disk-resident data whose size exceeds aggregate cluster memory. Such systems incur a significant performance penalty, however, on the important class of “groupwise set-valued analytics” (GSVA) queries in which the data is dynamically partitioned into groups and then a set-valued synopsis is computed for some or all of the groups. Key examples of synopses include top-k sets, bottom-k sets, and uniform random samples. Applications of GSVA queries include micro-marketing, root-cause analysis for problem diagnosis, and fraud detection. A naive approach to executing GSVA queries first reshuffles all of the data so that all records in a group are at the same node and then computes the synopsis for the group. This approach can be extremely inefficient when, as is typical, only a very small fraction of the records in each group actually contribute to the final groupwise synopsis, so that most of the shuffling effort is wasted. We show how to significantly speed up GSVA queries by slightly modifying the shared-nothing environment to allow tasks to occasionally access a small, common data structure; we focus on the Hadoop setting and use the “Adaptive MapReduce” infrastructure of Vernica et al. to implement the data structure. Our approach retains most of the advantages of a system such as Hadoop while significantly improving GSVA query performance, and also allows for incremental updating of query results. Experiments show speedups of up to 5x. Importantly, our new technique can potentially be applied to other shared-nothing systems with disk-resident data.

  PublisherDOI

Frequent coauthors

• Yanlei Diao

  University of Massachusetts Amherst

  13 shared

• Anna Liu

  8 shared

• Thanh T. L. Tran

  LinkedIn (United States)

  6 shared

• Yannis Sismanis

  IBM Research - Almaden

  4 shared

• Vuk Ercegovac

  Databricks (United States)

  4 shared

• Kai Zeng

  Kunming University of Science and Technology

  4 shared

• Charles Sutton

  4 shared

• Peter J. Haas

  Dortmund University of Applied Sciences and Arts

  4 shared

Labs

• Hospitality & Tourism ManagementPI

Awards & honors

• Nominee, Best Paper Award, 29th Graduate Education and Gradu…
• Best Paper Award, 28th Graduate Education and Graduate Resea…