High precision and reliable wind speed forecasting is a challenge for meteorologists. We used multiple nonparametric tree-based machine learning techniques, predicting the maximum at 10 m using selected convective weather variables. Analysis based on 127 storms from 2005 to 2013. The study evaluated two error models - Bayesian Additive Regression Trees (BART) Quantile Forests (QRF) compares them in terms of point estimates prediction intervals. model performances were different metrics...

This study presents a comprehensive investigation of multiple Artificial Intelligence (AI) techniques—decision tree, random forest, gradient boosting, and neural network—to generate improved precipitation estimates over the Upper Blue Nile Basin. All AI methods merged satellite atmospheric reanalysis datasets to error-corrected estimates. The accuracy model predictions was evaluated using 13 years (2000–2012) ground-based data derived from local rain gauge networks in Basin region. results...

We model the latent heats of crystallization and fusion in phase change materials with a unified heat change, ensuring energy conservation by coupling amorphous crystalline specific heats. demonstrate 2-D finite element simulations Ge2Sb2Te5 find that increases local temperature up to 180 K 300 nm × structures during crystallization, significantly impacting grain distributions. also show electrothermal 45 confined 10 mushroom cells higher predicted this nucleation probability at end reset...

Brain-implanted stimulators are revolutionizing treatment of many neurological and psychiatric diseases, but still rely on temporary batteries for power, which require periodic replacements. Battery-free, self-sustainable deep brain stimulation (DBS) devices remain an unsolved challenge. Herein, we report a self-sustainable, battery-free, intermittent DBS system. This device is enabled by interfacing high-performance bio-triboelectric nanogenerator (Bio-TENG) as energy harvester with...

We characterized resistance drift in phase change memory devices the 80 K to 300 temperature range by performing measurements on 20 nm thick, ∼70–100 wide lateral Ge 2 Sb Te 5 (GST) line cells. The cells were amorphized using 1.5–2.5 V pulses with ∼50–100 ns duration leading ∼0.4–1.1 mA peak reset currents resulting lengths between ∼50 and 700 nm. Resistance coefficients are calculated constant voltage starting as fast within a second after amorphization for 1 h duration. Drift ∼0.02 0.1...

We electrically characterized melt-quenched amorphized Ge2Sb2Te5 (GST) phase-change memory cells of 20 nm thickness, ∼66–124 width, and ∼100–600 length with without photoexcitation in the 80–275 K temperature range. The show distinctly different current–voltage characteristics low-field (≲19 MV/m), a clear response to optical excitation by red light, high-field (≳19 MV/m) regimes, very weak excitation. reduction carrier activation energy regime increases from ∼10 meV at 80 ∼50 150 (highest...

Resistance drift in amorphous Ge2Sb2Te5 is experimentally characterized melt-quenched line cells the range of 300 K to 125 and observed follow previously reported power-law behavior with coefficients 0.07 0.11 dark, linearly decreasing 1/kT. While these measured dark are similar commonly (∼0.1) at above room temperature, measurements under light show a significantly lower coefficient (0.05 illumination vs 0.09 150 K). Periodic on/off switching shows sudden decrease/increase resistance,...

We demonstrate a modified stochastic gradient ( Tiki-Taka v2 or TTv2 ) algorithm for deep learning network training in cross-bar array architecture based on ReRAM cells. There have been limited discussions arrays applications due to the challenges switching behavior of nonvolatile memory materials. is known overcome device non-idealities training. feasibility linear regression task using 1R and 1T1R devices. Using measured properties, we project performance long short-term (LSTM) with 78 K...

The rapid quench from melt associated with the reset operation in phase change memory is a suitable physical process to leverage programming variability for hardware security. We have experimentally characterized cell-to-cell Ge <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> Sb Te xmlns:xlink="http://www.w3.org/1999/xlink">5</sub> (GST) line cells using single voltage pulse moderate amplitude. observed that cell dimensions and...

We characterized resistance drift in phase change memory devices the 80 K to 300 temperature range by performing measurements on 20 nm thick, ~70-100 wide lateral Ge2Sb2Te5 (GST) line cells. The cells were amorphized using 1.5-2.5 V pulses with ~50-100 ns duration leading ~0.4-1.1 mA peak reset currents resulting lengths between ~50 and 700 nm. Resistance coefficients are calculated constant voltage starting as fast within a second after amorphization for 1 hour duration. Drift ~0.02 0.1...

We have computationally demonstrated logic function implementations using lateral and vertical multi-contact phase change devices integrated with CMOS circuitry, which use thermal cross-talk as a coupling mechanism to implement functions at smaller footprints. Thermal-crosstalk during the write operations is utilized recrystallize previously amorphized regions achieve toggle operations. Amorphized formed between different pairs of contacts are isolate read contacts. Typical expected...

Breach of security due to unauthorized access electronic hardware devices or chips has recently become a serious concern for the internet-connected daily activities. Imaging with electron microscopy is one invasive techniques used gain knowledge about chip layout and extract secret information by attackers. Automatic destruction disturbance key during such attacks are required ensure protection against these attacks. We have characterized caused programmed phase change memory (PCM) cells...

Hardware security primitives such as physical obfuscated keys (POKs) allow tamper-resistant storage of random based on manufacturing or variability. The output bits existing POK designs need to be first corrected due measurement noise using error correction methods and then de-correlated by privacy amplification processes. These additional requirements increase the hardware overhead reduce efficiency system. In this work, we propose an intrinsically reliable design capable generating...

Resistance drift and crystallization are critical concerns for accurate distinction between different states data retention in phase change memory (PCM), but their underlying physical mechanisms still not fully understood. In this work, we compared the resistance of suspended on-oxide amorphous Ge <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> Sb Te xmlns:xlink="http://www.w3.org/1999/xlink">5</sub> PCM line cells. We programmed 15...

Phase change memory (PCM) is an emerging high speed, density, endurance, and scalable nonvolatile technology which utilizes the large resistivity contrast between amorphous crystalline phases of chalcogenide materials such as Ge <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> Sb Te xmlns:xlink="http://www.w3.org/1999/xlink">5</sub> , (GST). In addition to being used a standalone memory, there has been growing interest in integration PCM...

We performed gradual programming of phase change memory (PCM) line cells with repeated pulses to transition from the crystalline amorphous state through various intermediate states and analyzed variability in trends for different cell shapes sizes. The PCM experienced sequences applied voltage either identical amplitudes or gradually increasing amplitudes. With amplitude pulses, large number (~1,000) were required reach a full reset partial set steps. ~20-30 sufficient fully Both approaches...

Physical Obfuscated Keys (POKs) allow tamper-resistant storage of random keys based on physical disorder. The output bits current POK designs need to be first corrected due measurement noise and next de-correlated since the original may not i.i.d. (independent identically distributed) also public helper information for error correction necessarily correlates bits.For this reason, include an interface and/or reinforcement, privacy amplification compressing a uniform bit string. We propose two...

Phase change memory (PCM) is a high speed, endurance, density non-volatile technology that utilizes chalcogenide materials such as Ge 2 Sb Te 5 (GST) can be electrically cycled between highly resistive amorphous and low resistance crystalline phases. The of the phase PCM cells increase (drift) in time following power law [1] , which increases window but limits implementation multi-bit-per-cell PCM. There has been number theories explaining origin drift – [4] mostly attributing it to...

Scaling of two dimensional six-contact phase change devices that can perform toggle logic operations is analyzed through 2D electrothermal simulations with dynamic materials modeling, integrated CMOS access circuitry. Toggle configurations are achieved a combination isolation some contacts from others using amorphous regions and coupling between different via thermal crosstalk. Use crosstalk as mechanism in multi-contact device the memory layer allows implementation analog routing digital at...

We observed resistance drift in 125 K - 300 temperature range melt quenched amorphous Ge2Sb2Te5 line-cells with length x width thickness = ~500 nm ~100 ~ 50 nm. Drift coefficients measured using small voltage sweeps appear to decrease from 0.12 +/- 0.029 at 0.075 0.006 K. The current-voltage characteristics of the amorphized cells 85 high-voltage (0 ~25 V) show a combination linear, low-field exponential and high-field conduction mechanisms, all which are strong functions temperature. very...

We stabilize resistance of melt-quenched amorphous Ge2Sb2Te5 (a-GST) phase change memory (PCM) line cells by substantially accelerating drift and bringing it to a stop within few minutes with application high electric field stresses. The acceleration is clearly observable at fields > 26 MV/m all temperatures (85 K - 300 K) independent the current forced through device, which strong function temperature. low-field (< 21 MV/m) I-V characteristics stabilized measured in 85 range fit well 2D...