Welcome to YODA – Open Digital Archive for CSEM
The YODA archive gives access to CSEM's publications, such as its annual reports and brochures. For technical papers such as scientific publications, bibliographic information is provided, along with the full paper where this is possible.
This comprehensive database is part of CSEM's Open Access Publishing policy. For further information refer to the YODA support guide or contact repository@csem.ch.
Recent Submissions
Item Neural Processing Unit (NPU): Accelerating neural network inference(2026)Artificial intelligence (AI) algorithms, such as neural networks, are fundamental to many advanced computer vision and signal processing capabilities, often exceeding human-level performance. These algorithms enable the automation of repetitive tasks by processing input data through millions of computations, achieving high accuracy and reliable performance. However, the significant memory and power requirements for processing these networks pose challenges for deployment in miniaturized, wearable, and other energy-constrained applications. Local processing of AI/ML algorithms within the end node, known as edge processing, can enhance the performance of many applications. However, unless a large battery size or frequent recharging can be afforded, embedded AI/ML processing is restricted to low complexity tasks or necessitates offloading to cloud processing. This offloading incurs the costs of energy-intensive radio communications, increased latency, and additional privacy concerns. Optimized and energy-efficient AI/ML chips address these challenges by accelerating neural network computations within a constrained power budget, thereby enhancing edge processing capabilities. CSEM’s next-generation neural processing unit (NPU) has been designed to address these challenges, enabling neural network edge processing in power- and energy-constrained embedded systems. The NPU is a standalone AI/ML accelerator IP that delivers state-of-the-art ML acceleration performance, optimized for embedded edge processing. With nearly 200x higher measured throughput (910GOP/s) and significantly increased efficiency (3.5TOPS/W) than its predecessor, CSEM’s latest AI/ML system-on-chip showcases the NPU’s cutting-edge performance for low-power AI chips.Item MFCC: Mel-Frequency Cepstrum Coefficients (MFCC) feature extractor(2026)The Mel-Frequency Cepstrum Coefficients (MFCC) core block is designed to extract features from voice signals with high efficiency. This technology is ideal for applications such as keyword spotting, speaker identification, and voice activity detection. This white paper provides a comprehensive overview of the MFCC core block architecture, specifications, and integration guidelines.Item Spectral Dynamics in Broadband Frequency Combs with Overlapping Harmonics(2025-11-21)Optical frequency combs and their spectra of evenly spaced discrete laser lines are essential to modern time and frequency metrology. Recent advances in integrated photonic waveguides enable efficient nonlinear broadening of an initially narrowband frequency comb to multi-octave bandwidth. Here, we study the nonlinear dynamics in the generation of such ultrabroadband spectra where different harmonics of the comb can overlap. We show that a set of interleaved combs with different offset frequencies extending across the entire spectrum can emerge, which can be arranged into a practically evenly spaced ultrabroadband frequency comb when the initial comb is offset-free.Item Dual-comb precision calibration of an astronomical Fabry–Pérot cavity(2025-12-01)Astronomical precision spectroscopy supports searches for exoplanets and may enable the direct observation of the Universe’s expansion. A key challenge in these efforts is the precise and long-term calibration of astronomical spectrographs. White light-illuminated Fabry–Pérot cavities, cross-calibrated with hollow-cathode lamps, are reliable calibrators. However, they exhibit temporal variability and nontrivial dispersion of the resonance frequencies, and the calibration lamps provide only sparse spectra with complex line shapes, which are also prone to aging. In contrast, laser frequency combs (“astrocombs”) can offer SI-referenced accuracy and precision, but the requirement of resolvable comb lines implies pulse repetition rates >10 GHz, which is challenging to implement. Here, we explore an alternative technique where lower-repetition rate dual-frequency comb lasers (comb lines unresolvable by the spectrograph) measure the modes of a 12 GHz Fabry–Pérot cavity across a spectral span of 30 THz (175–205 THz). This effectively transfers the comb’s accuracy to the cavity and enables precise near real-time tracking of the cavity modes. These measurements are of immediate relevance to currently used calibration techniques, and with further advances, this approach could potentially support spectrograph calibration at the level needed to search for Earth-like exoplanets.Item Progress Toward a Compact Two-Photon Rb Clock with Long-Term Stability Below 10-15(2025-05-12)Progress on the development of a compact twophoton Rb clock designed for timekeeping applications is presented. Current short-term and long-term limitations of the technology are discussed. Continuous operation during 4 months of an integrated prototype at the Swiss national metrology institute (METAS), measured against UTCr, is reported.
Communities in YODA
Select a community to browse its collections.