About

Elias Towe is a faculty member at Carnegie Mellon University, serving as a Professor in the Electrical and Computer Engineering and Material Science and Engineering departments. His work is associated with the Tepper School of Business, where he is involved in research and academic activities. Towe's expertise encompasses the intersection of business, technology, and analytics, contributing to the strategic vision of the Tepper School of Business. His research focuses on the integration of artificial intelligence and machine learning with business management, organizational behavior, and innovation, reflecting a data-informed, human-driven approach to solving complex problems. As a professor, he is engaged in advancing thought leadership and fostering practical applications of technology in business contexts.

Research topics

• Computer Science
• Artificial Intelligence
• Computer Security
• Quantum mechanics
• Engineering
• Physics
• Optoelectronics
• Neuroscience
• Materials science
• Theoretical computer science
• Biomedical engineering

Selected publications

• Photonic-digital hybrid artificial intelligence hardware architectures: at the interface of the real and virtual worlds

  Journal of Physics Photonics · 2025-09-10 · 1 citations

  articleOpen access

  Abstract Artificial intelligence (AI) technology has become an undeniable presence in society, as seen in the growing use of chatbots (e.g. ChatGPT) and recent Nobel Prize awards. However, the challenge of developing AI hardware architectures that efficiently integrate real-world analog signals with digital computational frameworks remains unresolved. This is particularly true for optical and neuromorphic computing systems. This work reports the implementation of an optical computing module in two hybrid AI architectures. The key element of the module is a photonic layer comprised of nanometer-scale spherical carbon dots. When excited with light-emitting diodes, the photonic layer exhibits a complex optical response. This layer is a feasible building block for two hybrid AI architectures: one based on optical processing and the other on the principles of neuromorphic computing. The modality of operation of the photonic layer is that it converts numerical input into a complex emission space, the results of which are processed using a Gaussian process model. Two systems based on this building block have been tested on a real-world dataset. They outperform conventional digital-only models, with coefficients of determination of r 2 = 0.90 and r 2 = 0.84 (training) and r 2 = 0.85 and r 2 = 0.87 (testing) for the optical and neuromorphic architectures, respectively. The systems discussed in this work serve as interfaces that bridge the real and virtual worlds. They offer exceptional optical properties, relative innocuity, low toxicity, stability, cost-effectiveness, seamless scalability, and robustness. The proposed architectures push the boundaries of advanced hybrid designs.

  PublisherDOI

• Development of regular vertical p-n junction on nanocrystalline PbTe film

  Journal of Applied Physics · 2025-03-14 · 3 citations

  articleOpen access

  Polycrystalline nanograined p-type PbTe films were obtained by electron gun-assisted vapor deposition on 100 μm thick amorphous substrates. This part of the study included the establishment and tuning of fabrication technology regimes in terms of the films' composition and crystallites arrangement optimal for having best structural properties, such as dominant texture, tiny-sized or absent voids, and small surface roughness. For this synthesis, we used components' composition Pb0.999Te1.001 bearing in mind that any excess Te builds up an acceptor center. Then, from the thus prepared p-type films, their n-type counterparts were obtained by ion implantation of zinc. At suitable conditions of the implantation process, the inversion of p-type to overall n-type material was experimentally shown and qualitatively explained. The structural and transport properties of both types of films were investigated, demonstrating their high integrity and a moderate effect of grain boundaries. Vertical p-n junction structures were prepared in the p-type films by a combination of proper masking and ion implantation. An electron beam-induced current technique was applied to directly portray the transition between p-and n-sides of the film and to assess the diffusion length of the minority charge carriers. The transition proves rather sharp spatially, which points to a well-defined p-n junction. Increasing the diffusion length of charge carriers of these structures compared to that in epitaxial films was discovered. A possible explanation of this effect and device applications of the developed structure are suggested.

  PublisherOA PDFDOI

• Advancing optoelectronic reservoir computing: enhancing performance through ultrafast neuromorphic hardware technologies

  Optics & Laser Technology · 2025-10-17 · 1 citations

  articleOpen access

  Reservoir computing is a neuromorphic architecture based on artificial neural networks. It has gathered significant attention due to its simplicity and efficiency in processing complex sequential data for real-world tasks. We propose an advanced optoelectronic reservoir computing system that uses a single nonlinear node comprised of a Mach-Zehnder interferometer, an optical delay line, and several high-bandwidth integrated optoelectronic components. This system shows efficient performance on benchmark tasks such as signal recognition with an accuracy of 100%, nonlinear channel equalization for generating reconstructed signals with symbol error rates of 10 −55 , and time-series predictions that reach normalized mean square errors in the order of 10 −2 .

  PublisherDOI

• Development of high-operation-temperature (up to 150 K) mid-wave infrared photodetectors based on <i>p</i>–<i>n</i> junctions in PbTe single crystals

  AIP Advances · 2024-08-01 · 2 citations

  articleOpen access

  One of the characteristic features of PbTe is an uncommon growth of bandgap with increasing temperature, which is quite opposite to the bandgap behavior of the semiconductors commonly used in electronics, for example, Si, Ge, GaAs, and InSb. This specificity allows one to increase the operating temperature of photodiodes fabricated using PbTe up to about 150 K. At the first stage of development, we prepared infrared (IR) photodiodes on the base of bulk single crystalline PbTe. To this end, the ingots with a diameter of about 40 mm were grown by the Czochralski technique. Then, the PbTe p–n junctions were fabricated by using indium donor diffusion to diffuse indium into the PbTe samples. Current–voltage and capacitance–voltage characteristics and spectral detectivity were measured over a wide temperature range and analyzed. The dark saturation current density at T = 100 K was of the order of 10−7 A/cm2. Finally, the unique solid-state multi-stage thermoelectric cooler operated at temperatures up to 150 K was developed. The present study would pave the way to creating a module for efficient photodetection in the mid-wave IR range combining two solid-state devices, namely, the p–n photodiode and thermoelectric cooler, while the latter supports the former.

  PublisherOA PDFDOI

• Luminescent Waveguides with Synaptic Properties for Photonic Artificial Neural Networks

  2024-01-01 · 1 citations

  article

  We replicate biological neurons and synapses, transmitting 0.2 Hz impulses through luminescent waveguides with adjustable features. This breakthrough has significant implications for neuromorphic engineering, providing valuable insights into neural networks technological applications and signal transmission.

  PublisherDOI

• Polycrystalline Films of Indium-Doped PbTe on Amorphous Substrates: Investigation of the Material Based on Study of Its Structural, Transport, and Optical Properties

  Materials · 2024-12-11 · 2 citations

  articleOpen access

  Nowadays, polycrystalline lead telluride is one of the premier substances for thermoelectric devices while remaining a hopeful competitor to current semiconductor materials used in mid-infrared photonic applications. Notwithstanding that, the development of reliable and reproducible routes for the synthesis of PbTe thin films has not yet been accomplished. As an effort toward this aim, the present article reports progress in the growth of polycrystalline indium-doped PbTe films and their study. The introduction foregoing the main text presents an overview of studies in these and closely related research fields for seven decades. The main text reports on the electron-beam-assisted physical vapor deposition of n-type indium-doped PbTe films on two different amorphous substrates. This doping of PbTe is unique since it sets electron density uniform over grains due to pinning the Fermi level. In-house optimized parameters of the deposition process are presented. The films are structurally characterized by a set of techniques. The transport properties of the films are measured with the original setups described in detail. The infrared transmission spectra are measured and simulated with the original optical-multilayer modeling tool described in the appendix. Conclusions of films' quality in terms of these properties altogether are drawn.

  PublisherOA PDFDOI

• Flexible optoelectrical neural implants

  2021

  • Computer Science
  • Artificial Intelligence
  • Computer Science

  With the advent of optical methods for stimulation and functional recording of neuronal activity in the brain, there is a growing need for fully flexible, ultracompact photonic devices for light delivery and light collection in brain tissue. In this paper, we will discuss our recent advances in designing a flexible optoelectronic neural implant platform that integrates passive and active optical components with electrical recording functionality. We leverage the exquisite optical and electrical insulation properties Parylene C, a biocompatible and flexible polymer to realize a fully functional optoelectrical neural interface.

  PublisherDOI

• GaN μLED Arrays in Parylene C Substrate for Flexible Implantable Optogenetics: Fabrication and Modeling

  Conference on Lasers and Electro-Optics · 2020-01-01 · 1 citations

  articleCorresponding

  Optoelectronic neural probes with (22 pm x 22 pm) Gallium-Nitride pLED arrays (up to 32 devices) on a Parylene C substrate are fabricated and analyzed using thermal and optical modeling to identify thermally-safe stimulation parameters.

  PublisherDOI

• On the Use of Quantum Entanglement in Secure Communications: A Survey

  arXiv (Cornell University) · 2020 · 15 citations

  • Computer Science
  • Computer Security
  • Computer Science

  Quantum computing and quantum communications are exciting new frontiers in computing and communications. Indeed, the massive investments made by the governments of the US, China, and EU in these new technologies are not a secret and are based on the expected potential of these technologies to revolutionize communications, computing, and security. In addition to several field trials and hero experiments, a number of companies such as Google and IBM are actively working in these areas and some have already reported impressive demonstrations in the past few years. While there is some skepticism about whether quantum cryptography will eventually replace classical cryptography, the advent of quantum computing could necessitate the use of quantum cryptography as the ultimate frontier of secure communications. This is because, with the amazing speeds demonstrated with quantum computers, breaking cryptographic keys might no longer be a daunting task in the next decade or so. Hence, quantum cryptography as the ultimate frontier in secure communications might not be such a far-fetched idea. It is well known that Heisenberg's Uncertainty Principle is essentially a "negative result" in Physics and Quantum Mechanics. It turns out that Heisenberg's Uncertainty Principle, one of the most interesting results in Quantum Mechanics, could be the theoretical basis and the main scientific principle behind the ultimate frontier in quantum cryptography or secure communications in conjunction with Quantum Entanglement.

  PublisherOA PDF

• High Density, Double-Sided, Flexible Optoelectronic Neural Probes With Embedded μLEDs

  Frontiers in Neuroscience · 2019-08-09 · 56 citations

  articleOpen access

  Optical stimulation and imaging of neurons deep in the brain require implantable optical neural probes. External optical access to deeper regions of the brain is limited by scattering and absorption of light as it propagates through tissue. Implantable optoelectronic probes capable of high-resolution light delivery and high-density neural recording are needed for closed-loop manipulation of neural circuits. Micro-light-emitting diodes (µLEDs) have been used for optical stimulation, but predominantly on rigid silicon or sapphire substrates. Flexible polymer neural probes would be preferable for chronic applications since they cause less damage to brain tissue. Flexible LED neural probes have been recently implemented by flip-chip bonding of commercially available µLED chips onto flexible substrates. Here, we demonstrate a monolithic design for flexible optoelectronic neural interfaces with embedded gallium nitride µLEDs that can be microfabricated at wafer-scale. Parylene C is used as the substrate and insulator due to its biocompatibility, compliance, and optical transparency. We demonstrate one-dimensional and two-dimensional individually-addressable µLED arrays. Our µLEDs have sizes as small as 22 µm × 22 µm in arrays of up to 32 µLEDs per probe shank. These devices emit blue light at a wavelength of 445 nm, suitable for stimulation of channelrhodopsin-2, with output power greater than 200 µW at 2 mA. Our flexible optoelectronic probes are double-sided and can illuminate brain tissue from both sides of the flexible probe. Recording electrodes are co-fabricated with µLEDs on the front- and backside of the optoelectronic probes for electrophysiology recording of neuronal activity from the volumes of tissue on the front- and backside simultaneously with bi-directional optical stimulation.

  PublisherOA PDFDOI

Recent grants

• Grapheotonics

  NSF · $150k · 2010–2013

Frequent coauthors

• Nobuo Saito

  Nagaoka University

  64 shared

• Jason C. S. Woo

  University of California, Los Angeles

  64 shared

• Zong Liang

  Victoria University of Wellington

  64 shared

• Karon E. MacLean

  University of British Columbia

  64 shared

• W. Bastiaan Kleijn

  64 shared

• Toma ́s Palacios

  University of British Columbia

  64 shared

• Lina J. Karam

  Arizona State University

  64 shared

• Devendra Pal

  36 shared