Training a ML model On An Intelligent Edge Device Using Less Than 256KB Memory


A new technical paper titled "On-Device Training Under 256KB Memory" was published by researchers at MIT and MIT-IBM Watson AI Lab. “Our study enables IoT devices to not only perform inference but also continuously update the AI models to newly collected data, paving the way for lifelong on-device learning. The low resource utilization makes deep learning more accessible and can have a bro... » read more

Research Bits: Oct. 10


Disposable water-activated battery Researchers at Empa developed a water-activated disposable paper battery that could be used in low-power, single-use disposable electronics such as smart labels for tracking objects, environmental sensors, and medical diagnostic devices. The battery is made of at least one cell measuring one centimeter squared and consisting of three inks printed onto a re... » read more

Week In Review: Manufacturing, Test


On Sunday, a 6.8-magnitude earthquake struck the southeast region of Taiwan, causing devastation. TSMC officials reported “no known significant impact for now.” Market research firm TrendForce arrived at a similar conclusion based on its analysis of individual fabs. The Biden administration announced appointment of the leadership team charged with implementing the US CHIPS and Science Ac... » read more

Research Bits: Aug. 30


Through glass vias Researchers from the Chinese Academy of Sciences (CAS) developed a Through Glass Via (TGV) process for 3D advanced packaging, which they say enables low transmission loss and high vacuum wafer-level packaging of high-frequency chips and MEMS sensors. TGV is a vertical interconnection technology applied in wafer-level vacuum packaging. The researchers found that it has goo... » read more

Technical Paper Round-Up: Aug 23


New technical papers added to Semiconductor Engineering’s library this week. [table id=46 /] Semiconductor Engineering is in the process of building this library of research papers. Please send suggestions (via comments section below) for what else you’d like us to incorporate. If you have research papers you are trying to promote, we will review them to see if they are a good fit for... » read more

Gemmini: Open-source, Full-Stack DNN Accelerator Generator (DAC Best Paper)


This technical paper titled "Gemmini: Enabling Systematic Deep-Learning Architecture Evaluation via Full-Stack Integration" was published jointly by researchers at UC Berkeley and a co-author from MIT.  The research was partially funded by DARPA and won DAC 2021 Best Paper. The paper presents Gemmini, "an open-source, full-stack DNN accelerator generator for DNN workloads, enabling end-to-e... » read more

Week In Review: Manufacturing, Test


The U.S. Commerce Department issued export controls on key technologies, including gallium oxide (Ga2O3) and diamond substrates, which are used at high voltages and temperatures, as well as EDA tools specifically developed for GAA FETs. It's not clear how this will impact EDA companies, because many of the tools that will be used for designing for GAA FETs already are in use today for finFETs. ... » read more

Techniques For Improving Energy Efficiency of Training/Inference for NLP Applications, Including Power Capping & Energy-Aware Scheduling


This new technical paper titled "Great Power, Great Responsibility: Recommendations for Reducing Energy for Training Language Models" is from researchers at MIT and Northeastern University. Abstract: "The energy requirements of current natural language processing models continue to grow at a rapid, unsustainable pace. Recent works highlighting this problem conclude there is an urgent need ... » read more

Security Research: Technical Paper Round-Up


A number of hardware security-related technical papers were presented at recent conferences, including the August 2022 USENIX Security Symposium and IEEE’s International Symposium on Hardware Oriented Security and Trust (HOST). Topics include side-channel attacks and defenses (including on-chip mesh interconnect attacks), heterogeneous attacks on cache hierarchies, rowhammer attacks and mitig... » read more

Microarchitectural Side-Channel Attacks and Mitigations on the On-Chip Mesh Interconnect


This new technical paper titled "Don't Mesh Around: Side-Channel Attacks and Mitigations on Mesh Interconnects" was presented by researchers at University of Illinois at Urbana-Champaign, MIT, and Texas Advanced Computing Center at the USENIX Security Symposium in Boston in August 2022. Abstract: "This paper studies microarchitectural side-channel attacks and mitigations on the on-chip mes... » read more

← Older posts Newer posts →