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:20260402T024509Z
LOCATION:3008\, Level 3
DTSTART;TZID=America/Los_Angeles:20250624T153000
DTEND;TZID=America/Los_Angeles:20250624T173000
UID:dac_DAC 2025_sess157@linklings.com
SUMMARY:Program, Debug, Accelerate: Software Innovations
DESCRIPTION:This session explores advances in sparse matrix computation, e
 mbedded AI, and system optimization. Topics include multimodal frameworks 
 and tensor compilers that enhance sparse matrix operations, on-device trai
 ning techniques that optimize neural network fine-tuning on constrained ha
 rdware, and AI-driven approaches for improving system security and reliabi
 lity. Additionally, novel storage optimizations and efficient data process
 ing methods will be presented. By integrating algorithmic innovations with
  hardware-aware strategies, these software innovations push the boundaries
  of performance, adaptability, and security in embedded computing.\n\nDroi
 dFuzz: Proprietary Driver Fuzzing for Embedded Android Devices\n\nEmbedded
  Android Devices have proliferated in many security-critical embedded scen
 arios, requiring sufficient testing to root out vulnerabilities. Due to An
 droid's architecture, which uses a Hardware Abstraction Layer (HAL) for ve
 ndor-specific driver implementations, traditional kernel testing techn...\
 n\n\nJianzhong Liu, Yuheng Shen, and Yifei Chu (Tsinghua University); Qian
 g Zhang (Hunan University); Heyuan Shi (Central South University); Wanli C
 hang (Hunan University); and Yu Jiang (Tsinghua University)\n-------------
 --------\nSSpMV: A Sparsity-aware SpMV Framework Empowered by Multimodal M
 achine Learning\n\nSparse Matrix-Vector Multiplication (SpMV) is an essent
 ial sparse operation in scientific computing and artificial intelligence.\
 nEfficiently adapting SpMV algorithms to diverse matrices and architecture
 s requires a framework capable of accurately recognizing sparse patterns a
 nd selecting the optimal...\n\n\nShengle Lin, Chubo Liu, and Yan Ding (Hun
 an University); Joey Zhou (Agency for Science, Technology and Research (A*
 STAR)); and Kenli Li, Wangdong Yang, and Xianhao He (Hunan University)\n--
 -------------------\nEnabling Data-Deduplication-Assisted Data Relocation 
 for Interlaced Magnetic Recording\n\nInterlaced Magnetic Recording (IMR) d
 rives have been regarded as promising hard disk drives (HDDs) to meet the 
 ever-growing storage demand in the post-AI era. Among various technologies
 , IMR drives achieve increased storage capacity by overlapping tracks in a
 n interlaced fashion; however, updating d...\n\n\nChen-Jui Tu and Shuo-Han
  Chen (National Yang Ming Chiao Tung University)\n---------------------\nA
 n Input-Aware Sparse Tensor Compiler Empowered by Vectorized Acceleration\
 n\nSparsity is widely prevalent in real-world applications, yet existing c
 ompiler optimizations and code generation techniques for sparse computatio
 ns remain underdeveloped. Sparse matrix-matrix multiplication (SpMM) is a 
 representative operator in sparse computations, whose performance is often
  limite...\n\n\nXianhao He, Haotian Wang, Jiapeng Zhang, and Wangdong Yang
  (Hunan University); Anthony Chronopoulos (The University of Texas at San 
 Antonio); and Kenli Li (Hunan University)\n---------------------\nSTREAM: 
 Spatiotemporal Similarity-based Efficient Approximate Median with Tunable 
 Granularity\n\nThe median (MED) is a crucial statistic for measuring the c
 entral tendency. However, exact MED computation remains costly, with even 
 state-of-the-art (SOTA) algorithms failing to meet (near) real-time proces
 sing demands. While approximate MED algorithm has arisen as a promising ca
 ndidate, existing a...\n\n\nFenfang Li and Huizhang Luo (Hunan University)
 , Weichen Liu (Nanyang Technological University), Anthony Chronopoulos (Th
 e University of Texas at San Antonio), and Kenli Li and Chubo Liu (Hunan U
 niversity)\n---------------------\nLocation is Key: Leveraging LLM for Fun
 ctional Bug Localization in Verilog Design\n\nIn Verilog code design, iden
 tifying and locating functional bugs is an important yet challenging task.
  Existing automatic bug localization methods have limited capabilities; th
 ey only suggest a set of potential buggy lines rather than precisely ident
 ifying the bug. Moreover, they depend on verificat...\n\n\nBingkun Yao and
  Ning Wang (City University of Hong Kong), Jie Zhou and Xi Wang (Southeast
  University), Hong Gao (Zhejiang Normal University), Zhe Jiang (Southeast 
 University), and Nan Guan (City University of Hong Kong)\n----------------
 -----\nEnabling On-Tiny-Device Model Personalization via Gradient Condensi
 ng and Alternant Partial Update\n\nOn-device training enables the model to
  adapt to user-specific data by fine-tuning a pre-trained model locally. A
 s embedded devices become ubiquitous, on-device training is increasingly e
 ssential since users can benefit from the personalized model without trans
 mitting data and model parameters to th...\n\n\nZhenge Jia, Yiyang Shi, Ze
 yu Bao, Zirui Wang, Xin Pang, Huiguo Liu, Yu Duan, Zhaoyan Shen, and Mengy
 ing Zhao (Shandong University)\n---------------------\nUnlocking a New Rus
 t Programming Experience: Fast and Slow Thinking with LLMs to Conquer Unde
 fined Behaviors\n\nTo provide flexibility and low-level interaction capabi
 lities, the ''unsafe'' tag in Rust is essential in many projects, but unde
 rmines memory safety and introduces Undefined Behaviors (UBs) that reduce 
 safety. Eliminating these UBs requires a deep understanding of Rust's safe
 ty rules and strong typ...\n\n\nRenshuang Jiang, Pan Dong, Zhenling Duan, 
 Yu Shi, Xiaoxiang Fang, Yan Ding, and Jun Ma (National University of Defen
 se Technology); Shuai Zhao (Sun Yat-sen University); and Zhe Jiang (Southe
 ast University)\n\nTopics: Systems\n\nTracks: SYS3: Embedded Software\n\nS
 ession Chairs: Hoeseok Yang (Santa Clara University) and Younghyun Kim (Pu
 rdue University)
END:VEVENT
END:VCALENDAR