Tag
Presents a novel estimation framework for D-vine copulas using gradient-based MLE and beam search for better global fit, and a localized anomaly detection method with uncertainty quantification via conformal prediction.
FlashTrie presents a GPU-accelerated constrained beam search for generative retrieval, using a succinct trie layout and cooperative CUDA kernels to reduce decoding latency and enable real-time serving at scale, achieving up to 24× speedup and a 0.71% revenue lift in a commercial search engine.
This paper proposes Retrieval Augmented Search (RAS), a blackbox adaptation method using retrieval-augmented search and atomic edit decomposition (AEGIS) to improve LLM-based program optimization for C++ and Python, achieving up to 2.06x improvement over prior methods.
This article explores using the gzip compression algorithm as a language model, demonstrating that compression algorithms can generate text by scoring candidate continuations based on compressed length, using beam search to produce output.
ThinkBooster is a unified framework for test-time compute scaling of LLM reasoning, providing a modular Python library, a performance-efficiency benchmark, an OpenAI-compatible proxy service, and a visual debugger. Empirical results on math and coding tasks demonstrate practical gains with quality-cost trade-offs.
This paper investigates how 1D coarse-to-fine token structures in autoregressive models improve test-time search efficiency compared to classical 2D grid tokenization. The authors show that such ordered tokens enable better test-time scaling and even training-free text-to-image generation when guided by image-text verifiers.