Recent publications
This article reports a high-density 3T1C single-finger (SF) embedded dynamic random access memory (eDRAM) compute-in-memory (CIM) macro. It features several techniques that enhance the memory density, the energy efficiency, and the throughput density, namely: 1) a high-density 3T1C SF-eDRAM cell with low-leakage retention (LLR) to improve the memory density significantly with a competitive retention time; 2) a bit-wise input-sparsity-aware (ISA) p-source input strategy for SF-eDRAM cell to save the energy dissipation of the eDRAM array; 3) a bit-significance-aware (BSA) analog-to-digital converter (ADC) to reduce the energy dissipation; and 4) a kernel-wise weight-update-and-refresh (KWUR) to improve the kernel-wise CIM utilization rate and the eDRAM-CIM macro throughput during weight update/refresh. The proposed 128-kb SF-eDRAM CIM macro prototyped in 28-nm CMOS exhibits a memory density of 2.28 Mb/mm
, reaches a peak throughput density of 18.7 TOPS/mm
, and a peak energy efficiency of 234.6 TOPS/W performing 8b operations.
This article introduces an ultra-low-power (ULP) active-RF tag for long-range event-driven Internet-of-Things (IoT) applications. The tag employs the following: 1) a complementary Class-B/-C voltage-controlled oscillator–power amplifier (VCO–PA) with an optimum-sized loop antenna and a co-designed transmit/receive (T/R) switch to maximize the efficiency of the transmitter (TX); 2) an antenna–transceiver (TRX) co-designed RF interface replacing the function of the off-chip matching network (MN) while enabling the frequency tunability and a 2-D wake-up scheme for the receiver (RX); and 3) a frequency-locked loop (FLL) plus injection locking calibration scheme for the TRX to guarantee both the frequency accuracy and tuning range. Prototyped in 65-nm CMOS, the TX achieves a 20.8
efficiency at
20.3-dBm equivalent isotropic radiated power (EIRP). The RX achieves a sensitivity of
60.4 dBm at a data rate of 1 kb/s while can operate from 790 to 925 MHz with 17.8-nW power consumption.
Compute-in-memory (CIM) is a promising approach for realizing energy-efficient deep neural network (DNN) accelerators. Previous CIM works focusing on uniform quantization (UQ) demonstrated a higher Multiply-accumulate (MAC) precision requirement to maintain DNN inferencing accuracy, resulting lower energy efficiency. The nonuniform quantization (NUQ) has proved to require lower precision than UQ, while the existing implementations are based on high precision digital lookup table (LUT) (e.g., 16-bit), leading to large energy and area overhead for multiplier. This work presents CLUT-CIM fabricated under 28-nm CMOS featuring: 1) a capacitance LUT (CLUT)-based NUQ MAC circuit with thermometer coding scheme for weight and input activation that avoids digital LUT and reduces the energy and area overhead; 2) a signed-channel training (SCT) method that reduces the switching activity of computation to improve the energy efficiency; 3) a dual-port 6T-SRAM array to enable simultaneously weight updating and CIM operations, enhancing the memory utilization and CIM throughput. Under 3-bit NUQ precision, the peak energy efficiency is 114.3 TOPS/W, and peak throughput density is 31.78 TOPS/mm
.
This article presents a direct time-of-flight (DToF) LiDAR with a current-integrating-based analog frontend (AFE), performing long-range (LR) and high-quality imaging. A current-integrating-based transimpedance amplifier (CI-TIA) facilitates a direct current-to-pulse conversion, which not only achieves superior transimpedance gain and noise performance but also adaptively compensates the background-light-induced direct current to alleviate the dynamic range (DR) obstacle in the LR LiDARs. Benefiting from the pulse manner output of the CI-TIA, a pulse position and width converter (PPWC) utilizes lightweight counters to provide the time and intensity information simultaneously, while the latter compensates for the walk error in time intrinsically. Furthermore, the intensity is used to adapt the accumulation window size to improve the imaging quality under high intensity. The prototype LiDAR receiver is fabricated in a 65-nm CMOS process, which achieves a consistent distance error of
1 cm and precision of
0.09% for up to 240-m measurements. It demonstrates a 256
192-pixel resolution imaging at 30 frames per second (fps) with the assistance of a scanning system.
Analog Voice Activity Detector (VAD) is a promising candidate for a power and cost-efficient solution for AIoT voice assistants. Regrettably, the PVT variation from the analog circuits and data misalignment from sensors limit the VAD accuracy with conventional backpropagation model-based training (BPMBT). This paper presents a forward-forward black box trainer (FFBBT) for analog VADs. It trains the analog circuit without knowing the circuit model or finding its gradient. Thus, it is insensitive to PVT variation and offset, achieving a measured VAD accuracy improvement of ~3% and an accuracy variation reduction of 5.6×. Moreover, a Tensor-Compressed Derivative-Free Optimizer (TCDFO) is also proposed to reduce the required memory for FFBBT by 1600×. The FFBBT with TCDFO is synthesized in 28 nm CMOS with a power of 512 nW and an area of 0.003 mm2.
Dielectric function models are essential for determining the optical constants of a substance as a function of photon energy using optical transmission, reflection or spectroscopic ellipsometry measurements. In this letter, we present an extended Tauc–Lorentz model tailored for amorphous materials with non-exponential band tails. Our method employs an exponential function with a polynomial argument to define the imaginary part of the dielectric function in the sub-gap region, with the polynomial order varying based on the complexity of sub-gap absorption features and the precision of the fitted experimental data. The real part of the dielectric function is obtained through the Kramers–Kronig relations as a sum of two components associated with interband and sub-gap transitions, allowing for the comparison of their contributions. These components are calculated analytically and numerically, simplifying the model’s implementation. We illustrate the model’s application by extracting the optical constants from the transmission spectrum of a hydrogenated silicon nitride thin film.
Dataflow management provides limited performance improvement to the transformer model due to its lesser weight reuse than the convolution neural network. The cosFormer reduced computational complexity while achieving comparable performance to the vanilla transformer for natural language processing tasks. However, the unstructured sparsity in the cosFormer makes it a challenge to be implemented efficiently. This brief proposes a parallel unstructured sparsity handling (PUSH) scheme to compute sparse-dense matrix multiplication (SDMM) efficiently. It transforms unstructured sparsity into structured sparsity and reduces the total memory access by balancing the memory accesses of the sparse and dense matrices in the SDMM. We also employ unstructured weight pruning cooperating with PUSH to further increase the structured sparsity of the model. Through verification on an FPGA platform, the proposed accelerator achieves a throughput of 2.82 TOPS and an energy efficiency of 144.8 GOPs/W for HotpotQA dataset with long sequences.
Tunneling magnetoresistive sensors with enhanced thermal resilience are being pursued for harsh environment applications. In the present work, we explore MnNi as a possible candidate for exchange bias in TMR sensor multilayers. We use ion beam assisted deposition for MnNi layer growth. A significant emergence of exchange bias field of
110 mT was obtained in MnNi/CoFe bilayers against the ion beam deposited counterpart. Additionally, we demonstrate for the first time in literature the compatibility with state-of-the-art sensor multilayers comprising synthetic antiferromagnets and MgO tunnel barriers. The optimized device shows a tunneling magnetoresistance ratio of 130 % at RT and a 30 % at 300
C corresponding to sensitivities of 17.0 %/mT and 9.5 %/mT with well defined parallel/antiparallel plateaus in the full temperature operation window.
This paper provides an overview of the global wave resource for energy exploration. The most popular metrics and estimators for wave energy resource characterization have been compiled and classified by levels of energy exploration. A review of existing prospective wave energy resource assessments worldwide is also given, and those studies have been collated and classified by continent. Finally, information about forty existing open sea wave energy test sites worldwide and their characteristics is depicted and displayed on a newly created global map. It has been found that wave power density is still the most consensual metric used for wave energy resource assessment purposes among researchers. Nonetheless, to accomplish a comprehensive wave resource assessment for exploitation, the computation of other metrics at the practicable, technical, and socio-economic levels has also been performed at both spatial and temporal domains. Overall, regions in latitudes between 40° and 60° of both hemispheres are those where the highest wave power density is concentrated. Some areas where the most significant wave power density occurs are in offshore regions of southern Australia, New Zealand, South Africa, Chile, the British Isles, Iceland, and Greenland. However, Europe has been the continent where most research efforts have been done targeting wave energy characterisation for exploitation.
Remote sensing technologies have the potential to support monitoring of floating plastic litter in aquatic environments. An experimental campaign was carried out in a large-scale hydrodynamic test facility to explore the detectability of floating plastics in ocean waves, comparing and contrasting different microwave and optical remote sensing technologies. The extensive experiments revealed that detection of plastics was feasible with microwave measurement techniques using X and Ku-bands with VV polarization at a plastic threshold concentration of 1 item/m² or 1–10 g/m². The optical measurements further revealed that spectral and polarization properties in the visible and infrared spectrum had diagnostic information unique to the floating plastics. This assessment presents a crucial step towards enabling the detection of aquatic plastics using advanced remote sensing technologies. We demonstrate that remote sensing has the potential for global targeting of plastic litter hotspots, which is needed for supporting effective clean-up efforts and scientific evidence-based policy making.
Quantum computation holds the promise of solving computational problems which are believed to be classically intractable. However, in practice, quantum devices are still limited by their relatively short coherence times and imperfect circuit-hardware mapping. In this work, we present the parallelization of pre-calibrated pulses at the hardware level as an easy-to-implement strategy to optimize quantum gates. Focusing on RZX gates, we demonstrate that such parallelization leads to improved fidelity and gate time reduction, when compared to serial concatenation. As measured by Cycle Benchmarking and Process Tomography, we reduce gate errors by half. We show that this strategy can be applied to other gates like the CNOT and CZ, and it may benefit tasks such as Hamiltonian simulation problems, amplitude amplification, and error-correction codes.
- B. Duval
- Abbas Abdolmaleki
- M. Agostini
- [...]
- M. Zurita
Tokamak à configuration variable (TCV), recently celebrating 30 years of near-continual operation, continues in its missions to advance outstanding key physics and operational scenario issues for ITER and the design of future power plants such as DEMO. The main machine heating systems and operational changes are first described. Then follow five sections: plasma scenarios. ITER Base-Line (IBL) discharges, triangularity studies together with X3 heating and N2 seeding. Edge localised mode suppression, with a high radiation region near the X-point is reported with N2 injection with and without divertor baffles in a snowflake configuration. Negative triangularity (NT) discharges attained record, albeit transient, βN ∼ 3 with lower turbulence, higher low-Z impurity transport, vertical stability and density limits and core transport better than the IBL. Positive triangularity L-Mode linear and saturated ohmic confinement confinement saturation, often-correlated with intrinsic toroidal rotation reversals, was probed for D, H and He working gases. H-mode confinement and pedestal studies were extended to low collisionality with electron cyclotron heating obtaining steady state electron iternal transport barrier with neutral beam heating (NBH), and NBH driven H-mode configurations with off-axis co-electron cyclotron current drive. Fast particle physics. The physics of disruptions, runaway electrons and fast ions (FIs) was developed using near-full current conversion at disruption with recombination thresholds characterised for impurity species (Ne, Ar, Kr). Different flushing gases (D2, H2) and pathways to trigger a benign disruption were explored. The 55 kV NBH II generated a rich Alfvénic spectrum modulating the FI fas ion loss detector signal. NT configurations showed less toroidal Alfvén excitation activity preferentially affecting higher FI pitch angles. Scrape-off layer and edge physics. gas puff imaging systems characterised turbulent plasma ejection for several advanced divertor configurations, including NT. Combined diagnostic array divertor state analysis in detachment conditions was compared to modelling revealing an importance for molecular processes. Divertor physics. Internal gas baffles diversified to include shorter/longer structures on the high and/or low field side to probe compressive efficiency. Divertor studies concentrated upon mitigating target power, facilitating detachment and increasing the radiated power fraction employing alternative divertor geometries, optimised X-point radiator regimes and long-legged configurations. Smaller-than-expected improvements with total flux expansion were better modelled when including parallel flows. Peak outer target heat flux reduction was achieved (>50%) for high flux-expansion geometries, maintaining core performance (H98 > 1). A reduction in target heat loads and facilitated detachment access at lower core densities is reported. Real-time control. TCV’s real-time control upgrades employed MIMO gas injector control of stable, robust, partial detachment and plasma β feedback control avoiding neoclassical tearing modes with plasma confinement changes. Machine-learning enhancements include trajectory tracking disruption proximity and avoidance as well as a first-of-its-kind reinforcement learning-based controller for the plasma equilibrium trained entirely on a free-boundary simulator. Finally, a short description of TCV’s immediate future plans will be given.
The formulation of the hybrid‐Trefftz stress element for plate bending is extended to the modelling of concentrated forces and moments, either as prescribed loads or as reactions at point supports. As the bending, torsion and shear fields are hypersingular, the flexibility matrix of the element involves the use of the finite part integration concept. In addition, it requires the confirmation of the positive‐definiteness of the flexibility under gross shape distortion. The tests illustrate the modelling of applied concentrated forces and moments and also the combination of boundary layer and point reaction effects. The results obtained are validated using converged solutions obtained with a stress‐based hybrid‐mixed element (HMS) and a displacement‐based mixed element (MITC).
Experimental investigations on the motion of rigid particles in microcirculation environments are still scarce owing to the three-dimensional (3D) motion of the particles and to the particle image masking due to the presence of the red blood cells (RBCs). Despite the recent progress on the 3D tracking of rigid particles in RBC flows with defocus particle tracking (DPT) methods, the problem of particle image masking remains to be solved. Here, we propose, test, and evaluate the use hemoglobin-free RBCs, also known as ghost RBCs, as a replacement for normal RBCs in experiments with rigid particles in microcirculation environments. We performed DPT measurements of a pressure-driven flow of normal and ghost RBC suspensions seeded with rigid particles at three different flow rates. We show that the quasi-transparent appearance of ghost RBCs, as a result of the lack of hemoglobin, eliminates the RBC-induced masking of the defocused particle images and allows to achieve the particle matching standards found in cell-free experiments. In fact, ghost RBC suspensions enable the tracking of the rigid particles across the entire height of the microchannel, which was not possible in normal RBC flows. On a fluid dynamic level, we show that ghost RBC suspensions provide similar conditions to normal RBCs in terms of the velocity of the rigid particles and the rigid particles exhibit similar lateral dynamics in both types of cell suspensions. In essence, the findings from this work demonstrate that ghost RBCs are a well-suited replacement for normal RBCs in experiments aiming at deciphering the motion of rigid particles in microcirculation environments.
Graphical abstract
Damage imaging algorithms are crucial for evaluating the condition of critical structures such as adhesively bonded joints. Particularly during service, baseline-free structural health monitoring is essential for robust and real-time evaluation. This paper proposes and investigates the impact of the shape of the damage intensity distribution and damage index on the damage imaging of composite lap joints using a baseline-free structural health monitoring system. This system comprises a parallel array of piezoelectric transducers attached to both sides of the lap joint for generating and receiving ultrasonic guided waves. Various features are extracted from the received signals to serve as damage indices, representing the peak amplitude and energy of the signals as well as the time-of-flight. Different shapes of damage intensity distribution, including elliptical, diamond, rectangular, and quadrilateral, are considered between pairs of sensors to investigate their effects on the total damage intensity distribution. To evaluate the impact of these parameters, a 2D correlation coefficient was employed to compare the results obtained from the baseline-free structural health monitoring system with the image containing actual defects. The results reveal that the time-of-flight was ineffective in providing high correlation and considering the signal's energy with quadrilateral shape achieved the highest correlation.
Technetium‐99 m (99mTc) remains the cornerstone of nuclear medicine for single photon emission computed tomography (SPECT) due to its widespread availability and chemical and physical features. Its multiple oxidation states allow for the design and production of radiopharmaceuticals with versatile properties, namely in terms of pharmacokinetic profile. 99mTc(V) is the most common oxidation state, but 99mTc(I) gained traction after the pioneering work of Alberto and colleagues, which resulted in the introduction of the organometallic core fac‐[99mTc(CO)3(H2O)3]⁺. This core is readily available from [99mTcO4]⁻ and displays three labile water molecules that can be easily swapped for ligands with different denticity and/or donor atoms in aqueous environment. This makes it possible to radiolabel small molecules as well as high molecular weight molecules, such as antibodies or other proteins, while assuring biological activity. Direct radiolabelling of those proteins with fac‐[99mTc(CO)3]⁺ under mild conditions is accomplished through incorporation of a polyhistidine tag (His‐tag), a commonly used tag for purification of recombinant proteins. This review aims to address the direct radiolabelling of His‐tagged macromolecules with fac‐[99mTc(CO)3]⁺ for development of molecular imaging agents and the impact of this technology in the discovery and development of imaging and/or therapeutic agents towards clinical application.
Despite all the fact-checking initiatives on news and social media aimed at countering misinformation, they remain insufficient to promptly address the wide array of misleading information disseminated by both news and social media outlets. Rather than attempting to identify or filter misleading information, this work advocates new tools for assisting online readers in identifying misinformation among the massive online content pushed every day through multiple platforms. We introduce DOMAIN, an article assessment resource bundle comprising a multidimensional indicator to categorize articles into different types (hard news, soft news, opinion, satire, and conspiracy), a set of explanatory metrics to help users understand the results, a tool for verifying the reliability of the article’s source, and a text summary of the assessment. This work also studies how DOMAIN tools impact online readers, specifically focusing on i) understanding the extent to which computer-generated assessments influence human perceptions of credibility; ii) evaluating the effectiveness of automatic article categorization in human assessment of credibility; and iii) identifying the most relevant explanatory metrics for promoting informed and critical consumption of information.
Customized fabrication of electrodes for electro-discharge machining (EDM) is considered as a useful emerging application of metal additive manufacturing technologies for building complex shaped tools in a short period of time. This approach of producing electrodes not only effectively addresses the design flexibility and complexity issues of the tools but also enables the use of various materials so as to ingrain the desired properties in them. The current study explores the production of laser powder bed fusion (LPBF) AlSi10Mg tools for machining of titanium alloy. Moreover, it explicitly assesses the machining performance of additively manufactured electrodes when the tools are subjected to a thin layer of copper coating. Three types of acidic baths are used for electroless coating of tools. The machining performance of the tools in terms of material removal rate, tool wear rate, surface crack density, surface roughness, white layer thickness, and microhardness of the white layer is compared for different electroless copper coating processes. It is observed that the coated tools significantly enhance material removal rate; however, uncoated LPBF tools exhibit superior surface morphology and surface roughness.
Institution pages aggregate content on ResearchGate related to an institution. The members listed on this page have self-identified as being affiliated with this institution. Publications listed on this page were identified by our algorithms as relating to this institution. This page was not created or approved by the institution. If you represent an institution and have questions about these pages or wish to report inaccurate content, you can contact us here.
Information
Address
Lisbon, Portugal