Visited UCSD, Samsung Research, and UCLA for invited talks and research collaboration on energy-efficient AI systems and large-scale AI optimization, alongside participating in NeurIPS 2025 with 5 accepted papers.
Participated in ICCAD 2025 in Munich and visited Politecnico di Torino, ETH Zurich, and IBM to discuss research collaboration on in-memory computing, neural architecture search, and efficient LLM serving.
IRIS Ph.D. student Hyeonsu Bang received the Excellence Award at the 2025 SKKU Graduate Student Paper Competition for a fault-tolerance scheme enabling reliable, efficient MLC ReRAM AI accelerators.
Jinju Kim’s CMU visit advances machine unlearning for speech/audio privacy, including NeurIPS 2025 AI4Music work.
Juhong Park’s Duke visit advances HW–SW co-optimization for LLM serving on DRAM-PIM/CIM NDP systems.
Presented our lab’s work on resource-efficient AI through SW–HW co-design at Duke, Cornell Tech, UIC, and UNC.
IRIS Lab undergraduate researchers Seokho Han and Seoyeon Yoon lead first-author paper at ICCV 2025, proposing MSQ for efficient on-device AI training.
We conducted a lab photo shoot, capturing profile and group photos of the professor and lab members.
Seminar by Prof. Yani Ioannou on SRigL and N:M structured sparsity with CPU/GPU speedups.
Rutgers’ Prof. Bokyung Kim presented DPIMA, a DRAM-based PIM accelerator for privacy-preserving ML (ISLPED 2025 Best Paper) at SKKU.
Prof. Jong Hwan Ko appointed as Research Unit 3 Lead in MSIT’s 6-year, 11B KRW AI Star Fellowship for next-generation AI talent.
MIT Tech Review featured our ICML 2025 voice-unlearning for zero-shot TTS, using TGU and spk-ZRF to curb deepfakes.
Prof. Jong Hwan Ko participates in Global Basic Research Laboratory project led by Prof. Kim Hwiyong, focusing on energy-efficient adaptive multipurpose coding.
KBS News featured IRIS Lab on Korea’s AI infrastructure challenges and our efficient GPU management efforts.
IRIS Lab contributes to Physical AI Center, developing HW/SW optimization for multi-sensor multimodal fusion in on-device AI systems.
Chanwook Hwang’s CityU visit advanced SNN-based spike detection for implantable BMI, with results presented at ISCAS 2025.
IRIS Lab undergraduate teams win Best Paper at IEIE AI Competition and Director’s Award at URP Showcase for dynamic neural network research.
MSIT’s YouTube series featured IRIS Lab’s on-device AI research and culture.
IRIS Lab’s Joochan Lee and Daniel Rho achieve Spotlight at ICLR 2024 and Highlight at CVPR 2024 for efficient 3D scene representation methods.
BK214-funded Hong Kong visit: CPAL 2024, HKUST 1D-CNN early detection, CityU neuromorphic/BMI talks.
IRIS Lab wins 1st place in Stage 2 of AI Grand Challenge auditory cognition track, securing MSIT Minister’s Award and 700M KRW research funding.
IRIS Lab takes first place in VisDrone Challenge object detection track, demonstrating multimodal AI expertise across audio, image, and video.
IRIS Lab places 3rd in auditory cognition track at MSIT’s AI Grand Challenge, securing 800M KRW follow-up research funding.
