BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
X-LIC-LOCATION:America/Los_Angeles
BEGIN:DAYLIGHT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
TZNAME:PDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
TZNAME:PST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260402T024534Z
LOCATION:Level 3 Lobby
DTSTART;TZID=America/Los_Angeles:20250622T180000
DTEND;TZID=America/Los_Angeles:20250622T190000
UID:dac_DAC 2025_sess261_RESEARCH1283@linklings.com
SUMMARY:HIVE: A Hierarchical Inverse Graph Construction Framework for Veri
 log Code Generation and Compilation Error Correction Using Large Language 
 Models
DESCRIPTION:Qingchen Zhai (Institute of Automation, Chinese Academy of Sci
 ences); Frank Qu (University of California, Santa Barbara); Hao Yu, Charle
 s Young, and LING LIANG (Peking University); Zhiwei Zhang (Institute of Au
 tomation, Chinese Academy of Sciences); Tao Xie (Peking University); and Y
 uan Xie (Hong Kong University of Science and Technology (HKUST))\n\nRecent
  advancements in large language models (LLMs) have shown significant poten
 tial in automating chip design, particularly in generating Verilog code fr
 om high-level specifications. While much of the existing research focuses 
 on end-to-end generation from specifications to code, there has been limit
 ed exploration of the linguistic style of module descriptions and the stru
 ctured representation of intermediate circuit forms. In this paper, we int
 roduce HIVE, a novel framework that improves Verilog code generation and d
 ebugging accuracy by integrating semantic style transfer, graph-structured
  reasoning and compilation error modification with retrieval augmented gen
 eration. The HIVE framework leverages GPT-3.5Turbo-FT and GPT-4Turbo, achi
 eving state-of-the-art performance on both the VerilogEvalv1 and Thakur's 
 benchmarks. Notably, HIVE-GPT3.5Turbo-FT eliminates syntax errors across a
 ll test cases in Thakur's benchmark. Furthermore, HIVE-GPT4Turbo achieves 
 a pass rate of 83.3% on the more challenging VerilogEvalv2 benchmark, demo
 nstrating its robustness in generating high-quality Verilog code.\n\nTrack
 s: DES5: Emerging Device and Interconnect Technologies\n\n
END:VEVENT
END:VCALENDAR
