Jiaxiang Yu

Xiaoxuan Liu, Jiaxiang Yu, Jongseok Park et al.

Speculative decoding (SD) has become a popular technique to accelerate Large Language Model (LLM) inference, yet its real-world effectiveness remains unclear as prior evaluations rely on research prototypes and unrealistically small batch sizes. We present, to our knowledge, the first systematic study of SD on a production-grade and widely deployed inference engine (vLLM), covering multiple SD variants ($n$-gram, EAGLE/EAGLE-3, Draft-Model, Multi-Token Prediction) across diverse workloads, model scales, and batch sizes. We analyze key factors governing SD performance, and quantify a theoretical upper bound on SD speedup. Our results show that verification by the target model dominates the execution, while acceptance length varies markedly across output token positions, requests, and datasets. Comparing measured performance with theoretical bounds reveals substantial gaps between observed and theoretical upper bounds, and we leverage this observation to highlight new research opportunities that our study opens up in improving SD.

Qiuyang Mang, Yufan Xiang, Hangrui Zhou et al.

Recent database systems have introduced semantic operators that leverage large language models (LLMs) to filter, join, and project over structured data using natural language predicates. In practice, these operators are combined with traditional relational operators, e.g., equi-joins, producing hybrid query plans whose execution cost depends on both expensive LLM calls and conventional database processing. A key optimization question is where to place each semantic operator relative to the relational operators in the plan: placing them earlier reduces the data that subsequent operators process, but requires more LLM calls; placing them later reduces LLM calls through deduplication, but forces relational operators to process larger intermediate data. Existing systems either ignore this placement question or apply simple heuristics without considering the full cost trade-off. We present Horrila, a plan-level optimizer for hybrid semantic-relational queries. Horrila reduces hybrid query planning to semantic filter placement via two equivalence-preserving rewrites. We prove that deferring all semantic filters to the latest possible position minimizes LLM invocations under function caching, but show that this can cause relational processing costs to dominate on complex multi-table queries. To balance LLM cost against relational cost, Horrila uses a dynamic-programming-based cost model that finds the placement minimizing their weighted sum. On 44 semantic SQL queries across five schemas and two benchmarks, Horrila achieves up to 1.5$\times$ speedup and 4.29$\times$ cost reduction while maintaining high output quality: an average F1 of 0.85 against the unoptimized baseline and 0.84 against human-annotated ground truth on SemBench. Overall, Horrila achieves a significant cost reduction while preserving the highest accuracy among six publicly available systems.

Tyler Griggs, Xiaoxuan Liu, Jiaxiang Yu et al.

Large language models (LLMs) are increasingly integrated into many online services, yet they remain cost-prohibitive to deploy due to the requirement of expensive GPU instances. Prior work has addressed the high cost of LLM serving by improving the inference engine, but less attention has been given to selecting the most cost-efficient GPU type(s) for a specific LLM service. There is a large and growing landscape of GPU types and, within these options, higher cost does not always lead to increased performance. Instead, through a comprehensive investigation, we find that three key LLM service characteristics (request size, request rate, SLO) strongly influence GPU cost efficiency, and differing GPU types are most cost efficient for differing LLM service settings. As a result, the most cost-efficient allocation for a given service is typically a mix of heterogeneous GPU types. Based on this analysis, we introduce Mélange, a GPU allocation framework that navigates these diverse LLM service characteristics and heterogeneous GPU option space to automatically and efficiently derive the minimal-cost GPU allocation for a given LLM service. We formulate the GPU allocation task as a cost-aware bin packing problem where GPUs are bins and items are slices of the service workload. Our formulation's constraints account for a service's unique characteristics, allowing Mélange to be flexible to support diverse service settings and heterogeneity-aware to adapt the GPU allocation to a specific service. Compared to using only a single GPU type, Mélange reduces deployment costs by up to 77% in conversational settings, 33% in document-based settings, and 51% in a mixed setting.

Jiaxiang Yu, Yiyang Liu, Ruiyang Fan et al.

In the facial expression recognition task, researchers always get low accuracy of expression classification due to a small amount of training samples. In order to solve this kind of problem, we proposes a new data augmentation method named MixCut. In this method, we firstly interpolate the two original training samples at the pixel level in a random ratio to generate new samples. Then, pixel removal is performed in random square regions on the new samples to generate the final training samples. We evaluated the MixCut method on Fer2013Plus and RAF-DB. With MixCut, we achieved 85.63% accuracy in eight-label classification on Fer2013Plus and 87.88% accuracy in seven-label classification on RAF-DB, effectively improving the classification accuracy of facial expression image recognition. Meanwhile, on Fer2013Plus, MixCut achieved performance improvements of +0.59%, +0.36%, and +0.39% compared to the other three data augmentation methods: CutOut, Mixup, and CutMix, respectively. MixCut improves classification accuracy on RAF-DB by +0.22%, +0.65%, and +0.5% over these three data augmentation methods.