SEFRQO combines LLMs, fine-tuning, and RAG to tackle the cold-start and generalization across workloads and DBMSes issues. This work is led by my students @hanwen_liu_usc and @ZhangQihan245, and in collaboration with @RyanMarcus.

Thrilled to share that our paper "SEFRQO: A Self-Evolving Fine-Tuned RAG-Based Query Optimizer" has been accepted to @sigmod 2026! Check its details in our preprint on https://t.co/Nq78BiKXWf and code on https://t.co/fY9vvpL7zt.

LIMAO is a lifelong learning framework that can be integrated with many learned query optimizers to boost their adaptability under shifting data/workloads. It supports modular plan decomposition, component reuse, and continual learning without retraining from scratch.

Excited to share that our paper “LIMAO: A Framework for Lifelong Modular Learned Query Optimization” has been accepted to @VLDBconf 2025. Check its details in our preprint on https://t.co/zy8TiNdHJi and code on https://t.co/ykV3qbQSAT.

Honored to receive the inaugural Google ML and Systems Junior Faculty Award! Thanks @Google for supporting my work at the intersection of machine learning and data-intensive systems. Full story: https://t.co/u9M52hP8br #Google #ML #DataSystems

Announcing the PhD and Graduate Studies Workshop & Student Research Competition (SRC) at ACM SIGSPATIAL 2025, which will be held on Monday, November 3rd, 2025, in Minneapolis, Minnesota! The submission deadline is **July 25, 2025**. For more details: https://t.co/WQrDvDXpKr.

Please spread the word among #SIGMOD2025 attendees. Looking forward to seeing you all in Berlin!

We received very encouraging reviews, highlighting that our demo can boost the research in the intersection of quantum computing and database systems, and make it approachable for more peers.

This is the "first" system prototype showing how quantum acceleration can be integrated into "real-time" query execution. We employ a hybrid quantum-classical approach to generate join order hints that can achieve up to 13x query execution speedup compared to vanilla PostgreSQL.

Thrilled to share that our demo on building a "quantum-augmented query optimizer", fully integrated into the PostgreSQL database engine, has been accepted to VLDB 2025 (@VLDBconf ). #QuantumComputing #Databases #VLDB2025

In this work, we are the first to use conformal prediction to verify, with formal guarantees, the learned query optimizer decisions and avoid disastrous plans. Also, we show how CP can proactively guide the query plan construction process to obtain verified learned plans upfront.

Excited to share that our paper “Conformal Prediction for Verifiable Learned Query Optimization” has been accepted to @VLDBconf 2025. Check its details in our preprint on arXiv:

This year, we are excited to host the 2nd edition of SoCal DB Day this Friday, May 9th @USC! The program is here: https://t.co/DHpRRKrVdi. A huge thank you to our incredible SoCal DB community, and a special thanks to our sponsors: @CSatUSC and @ebay.

The submission deadline for the Q-Data Workshop has been further extended to April 30, 2025 (11:59 PM PST). Plan to submit your work and spread the word! @SIGMODConf #SIGMOD2025

I’m very honored to be recognized as a Distinguished Reviewer at @SIGMODConf 2025!

The submission deadline for the Q-Data Workshop has been extended by one week. The new submission deadline is April 27, 2025. @SIGMODConf #SIGMOD2025

We took the first steps toward this goal and showcased a live demo of our proposed quantum-augmented query optimizer and its benefits in #Qubits25, running real TPC-H and IMDB benchmark queries. Stay tuned for more on realizing quantum-augmented database systems!

While formulating such optimization problems to show their applicability with quantum computing is an important first step, bridging the gap between theoretical formulations and practical, low-overhead integration into real-time query execution is a much bigger challenge.

We demonstrated how a hybrid quantum-classical solution can tackle NP-hard query optimization problems—such as finding optimal join orders—far more effectively than traditional classical approaches.

At #Qubits25, we showed that quantum computing can be "realized" as an accelerator within a real database system. I had the privilege to present our recent efforts in building a quantum-augmented query optimizer integrated into PostgreSQL. #QuantumOptimization #Databases #Dwave