Future Machines Research Articles

Experimental study on tritium breeding in water-cooled ceramic breeder blanket mock-up under D–T neutron irradiation conditions

The accurate assessment of tritium breeding parameters within fusion blankets is crucial for future magnet-confined fusion machines to realize fuel self-sufficiency. Such an assessment can be conducted using the simulation approach with nuclear data of high-fidelity and, most importantly, validated against experimental data. In this paper, we report neutronics experiment studies carried out on a mock-up of the water-cooled ceramic breeder blanket of the China fusion engineering test reactor (CFETR), irradiated at a deuterium–tritium (D–T) neutron source. The mock-up’s nuclear responses to 14 MeV neutrons, including tritium production rates (TPR) and neutron-induced reaction rates, are simulated and validated by experimental results. Redundant measurement techniques are used, including Li2TiO3 pellets for offline TPR measurements and the developed lithium glass detector with a significantly reduced size for online TPR measurements. The validation of the TPR value is complemented by the experimental evaluation of the neutron-induced reaction rates for Au and Zr foils. All experimental results are analyzed using the MCNP-4C code and FENDL-3.0 nuclear data library. The source term for the Monte Carlo simulation is built using a newly-developed method based upon the modeling of the depth profiling of tritium in the tritiated target. The experiments are in good agreement with the simulations; ratios of the calculation to experimental results (C/E) on TPR are found to be 0.97–1.08. The influence of the first wall tungsten armor on the mock-up nuclear responses is also studied.

Discrepancies in Stroke Distribution and Dataset Origin in Machine Learning for Stroke

BackgroundMachine learning algorithms depend on accurate and representative datasets for training in order to become valuable clinical tools that are widely generalizable to a varied population. We aim to conduct a review of machine learning uses in stroke literature to assess the geographic distribution of datasets and patient cohorts used to train these models and compare them to stroke distribution to evaluate for disparities. Aims582 studies were identified on initial searching of the PubMed database. Of these studies, 106 full texts were assessed after title and abstract screening which resulted in 489 papers excluded. Of these 106 studies, 79 were excluded due to using cohorts from outside the United States or being review articles or editorials. 27 studies were thus included in this analysis. Summary of reviewOf the 27 studies included, 7 (25.9%) used patient data from California, 6 (22.2%) were multicenter, 3 (11.1%) were in Massachusetts, 2 (7.4%) each in Illinois, Missouri, and New York, and 1 (3.7%) each from South Carolina, Washington, West Virginia, and Wisconsin. 1 (3.7%) study used data from Utah and Texas. These were qualitatively compared to a CDC study showing the highest distribution of stroke in Mississippi (4.3%) followed by Oklahoma (3.4%), Washington D.C. (3.4%), Louisiana (3.3%), and Alabama (3.2%) while the prevalence in California was 2.6%. ConclusionsIt is clear that a strong disconnect exists between the datasets and patient cohorts used in training machine learning algorithms in clinical research and the stroke distribution in which clinical tools using these algorithms will be implemented. In order to ensure a lack of bias and increase generalizability and accuracy in future machine learning studies, datasets using a varied patient population that reflects the unequal distribution of stroke risk factors would greatly benefit the usability of these tools and ensure accuracy on a nationwide scale.

NeMO-Net – Gamifying 3D Labeling of Multi-Modal Reference Datasets to Support Automated Marine Habitat Mapping

NASA NeMO-Net, The Neural Multimodal Observation and Training Network for global coral reef assessment, is a convolutional neural network (CNN) that generates benthic habitat maps of coral reefs and other shallow marine ecosystems. To segment and classify imagery accurately, CNNs require curated training datasets of considerable volume and accuracy. Here, we present a citizen science approach to create these training datasets through a novel 3D classification game for mobile and desktop devices. Leveraging citizen science, the NeMO-Net video game generates high-resolution 3D benthic habitat labels at the subcentimeter to meter scales. The video game trains users to accurately identify benthic categories and semantically segment 3D scenes captured using NASA airborne fluid lensing, the first remote sensing technology capable of mitigating ocean wave distortions, as well as in situ 3D photogrammetry and 2D satellite remote sensing. An active learning framework is used in the game to allow users to rate and edit other user classifications, dynamically improving segmentation accuracy. Refined and aggregated data labels from the game are used to train NeMO-Net’s supercomputer-based CNN to autonomously map shallow marine systems and augment satellite habitat mapping accuracy in these regions. We share the NeMO-Net game approach to user training and retention, outline the 3D labeling technique developed to accurately label complex coral reef imagery, and present preliminary results from over 70,000 user classifications. To overcome the inherent variability of citizen science, we analyze criteria and metrics for evaluating and filtering user data. Finally, we examine how future citizen science and machine learning approaches might benefit from label training in 3D space using an active learning framework. Within 7 months of launch, NeMO-Net has reached over 300 million people globally and directly engaged communities in coral reef mapping and conservation through ongoing scientific field campaigns, uninhibited by geography, language, or physical ability. As more user data are fed into NeMO-Net’s CNN, it will produce the first shallow-marine habitat mapping products trained on 3D subcm-scale label data and merged with m-scale satellite data that could be applied globally when data sets are available.

MFIX-Exa: A path toward exascale CFD-DEM simulations

MFIX-Exa is a computational fluid dynamics–discrete element model (CFD-DEM) code designed to run efficiently on current and next-generation supercomputing architectures. MFIX-Exa combines the CFD-DEM expertise embodied in the MFIX code—which was developed at NETL and is used widely in academia and industry—with the modern software framework, AMReX, developed at LBNL. The fundamental physics models follow those of the original MFIX, but the combination of new algorithmic approaches and a new software infrastructure will enable MFIX-Exa to leverage future exascale machines to optimize the modeling and design of multiphase chemical reactors.

Computational Infrared Spectroscopy of 958 Phosphorus-Bearing Molecules

Phosphine is now well-established as a biosignature, which has risen to prominence with its recent tentative detection on Venus. To follow up this discovery and related future exoplanet biosignature detections, it is important to spectroscopically detect the presence of phosphorus-bearing atmospheric molecules that could be involved in the chemical networks producing, destroying or reacting with phosphine. We start by enumerating phosphorus-bearing molecules (P-molecules) that could potentially be detected spectroscopically in planetary atmospheres and collecting all available spectral data. Gaseous P-molecules are rare, with speciation information scarce. Very few molecules have high accuracy spectral data from experiment or theory; instead, the best current spectral data was obtained using a high-throughput computational algorithm, RASCALL, relying on functional group theory to efficiently produce approximate spectral data for arbitrary molecules based on their component functional groups. Here, we present a high-throughput approach utilizing established computational quantum chemistry methods (CQC) to produce a database of approximate infrared spectra for 958 P-molecules. These data are of interest for astronomy and astrochemistry (importantly identifying potential ambiguities in molecular assignments), improving RASCALL's underlying data, big data spectral analysis and future machine learning applications. However, this data will probably not be sufficiently accurate for secure experimental detections of specific molecules within complex gaseous mixtures in laboratory or astronomy settings. We chose the strongly performing harmonic ωB97X-D/def2-SVPD model chemistry for all molecules and test the more sophisticated and time-consuming GVPT2 anharmonic model chemistry for 250 smaller molecules. Limitations to our automated approach, particularly for the less robust GVPT2 method, are considered along with pathways to future improvements. Our CQC calculations significantly improve on existing RASCALL data by providing quantitative intensities, new data in the fingerprint region (crucial for molecular identification) and higher frequency regions (overtones, combination bands), and improved data for fundamental transitions based on the specific chemical environment. As the spectroscopy of most P-molecules have never been studied outside RASCALL and this approach, the new data in this paper is the most accurate spectral data available for most P-molecules and represent a significant advance in the understanding of the spectroscopic behavior of these molecules.

Liquid Metal‐Enabled Soft Logic Devices

The controllable manipulation of gallium‐based liquid metal (LM) droplet presents a promising option toward the goal of making fully soft machines, which features an autonomy via external administration apparatus. While most current endeavors mainly concentrate on LM‐assisted actuations for soft robots, it is noteworthy that the peculiar behavior and LM droplet circuits thus enabled can be adopted as intelligent elements. Such characteristics will pave a way to innovate conventional logic devices and emancipate soft robots from the existing rigid machinery. Herein, LM‐based soft logic systems are described which are capable of regulating soft machines on macroscale. The core element is a LM‐based switch (LMS) with two specific types, both utilizing the LM droplet as a conducting medium. Soft logic gates of AND, OR, NOT, NAND, and NOR thus fabricated are demonstrated to generate electrical outputs as an integration of random inputs, and these signals can be adapted for electronic devices. For practical illustration, a logic accumulator is further constructed which can effectively regulate a pneumatic soft gripper within four states of inflation. The present LM‐enabled logic devices are expected to offer a feasible strategy to manipulate future soft machines without the participation of traditional rigid controlling elements.

Identifying Key Determinants of Childhood Obesity: A Narrative Review of Machine Learning Studies.

Machine learning is a class of algorithms able to handle a large number of predictors with potentially nonlinear relationships. By applying machine learning to obesity, researchers can examine how risk factors across multiple settings (e.g., school and home) interact to best predict childhood obesity risk. In this narrative review, we provide an overview of studies that have applied machine learning to predict childhood obesity using a combination of sociodemographic and behavioral risk factors. The objective is to summarize the key determinants of obesity identified in existing machine learning studies and highlight opportunities for future machine learning applications in the field. Of 15 peer-reviewed studies, approximately half examined early childhood (0-24 months of age) determinants. These studies identified child's weight history (e.g., history of overweight/obesity or large increases in weight-related measures between birth and 24 months of age) and parental overweight/obesity (current or prior) as key risk factors, whereas the remaining studies indicated that social factors and physical inactivity were important in middle childhood and late childhood/adolescence. Across age groups, findings suggested that race/ethnic-specific models may be needed to accurately predict obesity from middle childhood onward. Future studies should consider using existing large data sets to take advantage of the benefits of machine learning and should collect a wider range of novel risk factors (e.g., psychosocial and sociocultural determinants of health) to better predict childhood obesity. Ultimately, such research can aid in the development of effective obesity prevention interventions, particularly ones that address the disproportionate burden of obesity experienced by racial/ethnic minorities.

Scaling of impurity fractions for divertor detachment in high-density high-power operation scenarios

For future reactor designs and for planning of operational scenarios in the present and future machines, impurity fractions in the tokamak divertor plasma, c Z , is an essential input for predictive scalings of divertor detachment. A new quantitative scaling law for c Z , averaged over the scrape-off layer (SOL), required to attain detachment is developed for high power H-mode plasmas operated at high densities close to the density limit. It is based on a simple SOL radiation model which uses the combination of the empirical scaling laws for the H-mode power threshold and the separatrix density limit imposed by MHD instabilities. Additionally, it assumes, in agreement with experimental observations that the width of the power conducting layer outside of the separatrix scales approximately with the ion poloidal gyro-radius. The derived expression for c Z scales strongly with toroidal magnetic field, B, major radius R, the factor of access of the power flow through the separatrix over that required for the L–H transition, f LH, and isotope mass A: . Estimates of required impurity fractions for divertor detachment for an number of impurity species (N2, Ne and Ar) in future tokamak reactors ITER and DEMO are made in the paper.

Investigation of Hydrogen Glow Discharge Cleaning Side Effects on Tungsten

Hydrogen glow discharge cleaning (H-GDC) is a routine conditioning procedure for the modern tokamaks and the future fusion machines, including ITER. Due to the low energy of hydrogen ions in glow discharge plasmas, the probability of any considerable damage to the plasma-facing components was mainly ignored. In this work, Tungsten, as the primary candidate for the plasma-facing materials in tokamaks, is considered for studies regarding effects of the H-GDC procedure on the material during a routine vacuum vessel conditioning in Damavand tokamak. After performing routine H-GDC using pure hydrogen, the formation of loosely attached nanostructured bundles (NSBs) on the surface of the tungsten samples was observed. The presence of the NSBs, which can be a source of dust, would be significant due to their possible effects on the functionality of the future fusion plasma devices. The NSBs are observed on the tungsten samples with the surface temperature of <370 K after 2.5–4 hours of hydrogen ion bombardment with incident energy of ≈120 eV and fluence of ≈(2–3.5) × 1022 m–2. The surface modifications of specimens exposed to H‑GDC were examined using a material probe experiment and several surface analysis techniques such as SEM, EDX, XRD, and ERDA. Therefore, the formation of loose nanostructures on the wall of plasma confinement vessels due to H-GDC draws attention to probable damaging effects of this phenomenon upon functionality and outcomes of tokamaks.

Influence Function Measurement Technique Using the Direct and Indirect Piezoelectric Effect for Surface Shape Control in Adaptive Systems

Since the introduction of adaptive systems for corrective measures, static and dynamic disturbances have been reduced through the manipulation of surface shape in a well-controlled manner. One important application where adaptive systems are highly needed is in the future photo-lithography machines, particularly the wafer table where disturbances affect overlay and focus performance. To reduce overlay and focus errors, a dense array of actuators must be integrated into the wafer table to apply both in- and out-of-plane corrective deformations to the wafer surface. To realize a certain wafer shape, influence functions are linearly superimposed. Accurate models of the influence functions result in an accurate prediction of the final wafer shape. To reduce calibration errors, the influence function of every actuator needs to be determined. Here, we propose a technique to rapidly measure the influence function: the influence function measurement (IFM) technique. Using the actuate-sense property of piezoelectric materials, the influence function is determined by activating one piezo-actuator and measuring the charge induced on the neighboring piezo-actuators. The measured charges are compared to the results of finite element simulations and the absolute difference of 0.3 &#x0025; is reported for the inter-actuator charge coupling parameters. This clearly indicates the potential of the proposed technique. <i>Note to Practitioners</i>&#x2014;A critical step during the integrated-circuit lithographic fabrication process is the exposure. A particular practical challenge in this step is to correct for the mismatch between the optical image plane and the wafer surface, caused by wafer deformations. The mismatch negatively impacts the overlay and focus performance. To reduce the mismatch, actuators can be embedded into the wafer table to counteract wafer deformations, where the overall wafer surface shape is a superposition of each actuator&#x2019;s influence function. As part of the calibration for the embedded actuators, their influence functions are measured and stored. However, errors arise when influence functions are not measured for every actuator. Additionally, for different wafers, the influence functions vary hence the calibration should be done per wafer. This calibration process takes time and should not influence throughput. Our paper presents a viable approach that reduces the calibration time of active wafer tables, which is also relevant for other adaptive systems such as deformable mirrors. Since the technique requires the actuators to be mechanically coupled, it is limited to continuous surfaces. The proposed technique is numerically tested and experimentally verified to demonstrate performance.

Current and future machine used on blood purification in critical care

Current and future machine used on blood purification in critical care

The Fury Archives: Female Citizenship, Human Rights, and the International Avant-Gardes by Jill Richards

Reviewed by: The Fury Archives: Female Citizenship, Human Rights, and the International Avant-Gardes by Jill Richards Linda Roland Danil, Dr. (bio) Jill Richards, The Fury Archives: Female Citizenship, Human Rights, and the International Avant-Gardes (Columbia University Press 2020), ISBN: 9780231197113 (paperback), 344 pages. Already at the outset of her rich and varied book, Jill Richards makes her intent clear: her aim is to “assemble a transatlantic archive of female citizenship.”1 Richards’ analysis is contextualized in wider narratives of women’s rights [End Page 831] during the period generally referred to as feminism’s first wave during the late nineteenth and early twentieth centuries. Moreover, rather than focus on female citizenship as a priori constituted through shared histories of exclusion and as being underpinned by histories and experiences of shared pain and suffering—Richards focuses on the female/woman as a subject of action. Richards therefore considers women arsonists, industrial saboteurs, queer resistance cells, and suffrage rioters, among others. The thesis of Richards’ book is that women’s rights movements and the historical avant-gardes became entangled in “an often vexed but hugely influential relationship of reciprocal construction.”2 In her analysis, Richards further draws on a wide range of sources, and includes parliamentary inquiries, police reports, mug shots, and propaganda booklets, among others. Specifically, Richards’ book is situated within the historical timeframe when first wave feminism and institutional human rights emerged at the same time as Dada and surrealism, and asks—how can these currents be understood alongside each other and as part of the same historical period? More specific still, Richards focuses on the little-known but significant entanglement between women’s rights movements and the international avant-gardes, with that interweaving having partly arisen through a mutual dependence on a socialist international that provided financial backing for the distribution of radical newspapers, journals and pamphlets across the Atlantic. The individuals (not always women) that populate the book’s pages are endlessly fascinating, intelligent, creative, and principled. In the analysis of the intersection between art and politics, Richards writes about women such as Paulette Nardal, who produced a black feminist periodical and served as an area expert for the newly founded United Nations; Leonora Carrington, the surrealist painter, was a refugee a number of times over and fled to Mexico in fear of reprisal for her involvement in anti-government protest meetings; Til Brugman joined the Dutch resistance; the Dada artist Hannah Höch was an advocate for birth control reform and her work was condemned as ‘degenerate art’ by the Nazis; the novelist Rebecca West rioted with the suffragettes. Other individuals include the gender non-conforming artist Claude Cahun and their partner, Marcel Moore— who were both sent to a prisoner of war camp for their participation in the French resistance. Not all of the individuals in question were admirable or even remotely likable, however—think of the despicable Marie Huot, the nineteenth century secretary of the League Against Vivisection—who claimed to have disposed of new-born puppies by tying a heavy stone around their necks and then tossing them into a body of water. In another anecdote, Huot apparently imparted the advice that housecat populations could be controlled by directly taking the kittens away from the mother and drowning them in a pot of water closed with a heavy lid. Huot allegedly argued that the same could be done for unwanted human infants, and further advocated for the entire extinction of humanity through the use of birth control. In an engaging discussion in which Huot is first brought up, Richards discusses the notion of proletarian “birth strikes” or womb strikes, in which women would deploy birth control as a form of strike against the production of people to feed [End Page 832] to future war machines and the industrial production of commodities. Through this view, “the limitation of reproduction would deprive capitalism of workers and soldiers to exploit, eventually resulting in a world-wide revolution.”3 Here, workers took control of (self)reproduction in a biopolitical form of resistance against the state, and rather than reproductive rights being framed through the language of the human right to privacy to be granted by the...

A Q-Learning-Based Adaptive MAC Protocol for Internet of Things Networks

In Internet of Things (IoT) applications, sometimes the quality of service (QoS) of throughput for transmitting video or the QoS of bounded delay for control of a sensor node is required. A traditional contention-based medium access control (MAC) protocol cannot meet the adaptive traffic demands of these networks and confers delay-related constraints. Q-learning (QL) is one of the reinforcement learning (RL) mechanisms and can potentially be the future machine learning scheme for spectrum MAC protocols in IoT networks. In this study, a QL-based MAC protocol is proposed to facilitate adaptive adjustment of the length of the contention period in response to the ongoing traffic rate in IoT networks. The novelty of QL-based MAC lies in its use of RL to dynamically adjust the length of the contention period according to the traffic rate. The QL-based MAC will solve the models without additional input information to adapt to environmental variations during training. We confirm that the proposed QL-based MAC protocol with node contention is robust. In addition, we showed that our proposed QL-based MAC protocol has higher system throughput, lower end-to-end delay, and lower energy consumption in MAC contention than those of contention-based MAC protocols.

Combination of network and molecule structure accurately predicts competitive inhibitory interactions

Mining of metabolite-protein interaction networks facilitates the identification of design principles underlying the regulation of different cellular processes. However, identification and characterization of the regulatory role that metabolites play in interactions with proteins on a genome-scale level remains a pressing task. Based on availability of high-quality metabolite-protein interaction networks and genome-scale metabolic networks, here we propose a supervised machine learning approach, called CIRI that determines whether or not a metabolite is involved in a competitive inhibitory regulatory interaction with an enzyme. First, we show that CIRI outperforms the naive approach based on a structural similarity threshold for a putative competitive inhibitor and the substrates of a metabolic reaction. We also validate the performance of CIRI on several unseen data sets and databases of metabolite-protein interactions not used in the training, and demonstrate that the classifier can be effectively used to predict competitive inhibitory interactions. Finally, we show that CIRI can be employed to refine predictions about metabolite-protein interactions from a recently proposed PROMIS approach that employs metabolomics and proteomics profiles from size exclusion chromatography in E. coli to predict metabolite-protein interactions. Altogether, CIRI fills a gap in cataloguing metabolite-protein interactions and can be used in directing future machine learning efforts to categorize the regulatory type of these interactions.

Use of Machine Learning in Stock Market Prediction

Objective: The objective of this study was to examine and determine future directions in regard to future machine learning techniques based on the review of the current literature.&#x0D; Methodology: A systematic review has been used to review the current trends from the peer-reviewed journal articles in the past twenty years. For this study, four categories have been categorized, the use of neural networks, support vector machines, the use of a genetic algorithm, and the combination of hybrid techniques. Studies in each of these categorize have been evaluated.&#x0D; Finding: Firstly, there is a strong link between machine learning methods and the prediction problems they are associated with. The second conclusion that we can conclude from this review is that past studies need to improve its generalizability results. Most of the studies that have been reviewed in this analysis has only used the machine learning systems through the use of one market or during only a one time period without taking into consideration whether the system would be adaptable in other situations and conditions. Limitations, future trends, as well as policy implications have been defined.

Design of helical gear with carbon reinforced EN36 steel for two stage constant mesh gearbox weight reduction

Engineering design is an iterative process that requires to be dealt with all feasible design solutions in order to arrive at a desired objective. Gear drives are considered as the critical components in a mechanical power transmission system. It is possible that gears will predominate as the most effective means of transmitting power in future machines due to their high degree of reliability and compactness. With the advancing technology in recent years, the machines are demanding varying speed and load. Over the years, research has been done predominantly on gears to minimize or to reduce the failures and for optimal design. The main aim of this work is to design and analyse the hybrid composite gears for two stage constant mesh helical gearbox and to compare the stress and deformation results with and without the inclusion of Carbon Fibre Reinforced (CFR) material. The objective of this research work is to reduce weight and rotational inertia of the entire assembly by using Hybrid Helical Gears. Carbon Fibre Standard UT Prepreg was selected for this study because of its high strength to weight ratio. By comparing the weights of EN36 Steel Gear and Hybrid Gear, a 25.3% reduction in weight was observed. This research work illuminates the replacement of EN36 alloy steel gear in terms of stresses and deformation with an added advantage of reduced weight and rotational inertia of the helical gear by using EN36/Carbon Fibre Composite material while considering the safety of the component in the application of two-stage constant mesh gearbox.

Active friction control in lubrication condition using novel metal morphing surface

We propose a new active friction control method with a morphing surface using a unique diaphragm structure consisting of 60-μm thin metal film. In poor lubrication, the friction coefficient dropped significantly from 0.19 to 0.03 when the surface changed from concave to convex. Optical in situ observations revealed that the oil collected at the convexity due to capillary force and meniscus. In addition, results of electric characteristic measurements revealed that the solid contact area decreased with the convexity. In conclusion, the reduction of solid contact due to improvement of lubrication condition was likely the friction reduction mechanism of the convex shape. Because active friction control realizes minimal friction loss and multifunctionality, the morphing surface can contribute to future machine development.

Artificial intelligence and legal liability: towards an international approach of proportional liability based on risk sharing

ABSTRACT This paper critically examines the allocation of liability when autonomous artificial intelligence (AI) systems cause accidents. Problems of applying existing principles of legal liability in AI environment are addressed. This paper argues that the sharing of risk as a basis for proportionate liability should be a basis for a new liability regime to govern future autonomous machines. It is argued that this approach favors the reality of parties’ consent to taking the risk of unpredictable AI behavior over the technicality of existing principles of legal liability. The suggested approach also encourages transparency and responsible decisions of developers and owners of AI systems. A flowchart to clarify possible outcomes of applying the suggested approach is provided. The paper also discusses the need for harmonization of national laws and international cooperation regarding AI incidents crossing national borders to ensure predictability of legal rules governing the liability ensuing from AI applications.

Edge Learning with Timeliness Constraints: Challenges and Solutions

Future machine learning (ML) powered applications, such as autonomous driving and augmented reality, involve training and inference tasks with timeliness requirements and are communication- and computation-intensive, which demands the edge learning framework. The real-time requirements drive us to go beyond accuracy for ML. In this article, we introduce the concept of timely edge learning, aiming to achieve accurate training and inference while minimizing the communication and computation delay. We discuss key challenges and propose corresponding solutions from data, model, and resource management perspectives to meet the timeliness requirements. In particular, for edge training, we argue that the total training delay rather than rounds should be considered, and propose data or model compression, and joint device scheduling and resource management schemes for both centralized training and federated learning systems. For edge inference, we explore the dependency between accuracy and delay for communication and computation, and propose dynamic data compression and flexible pruning schemes. Two case studies show that the timeliness performance, including the training accuracy under a given delay budget and the completion ratio of inference tasks within deadline, are highly improved with the proposed solutions.

Return on Investment in Machine Learning: Crossing the Chasm between Academia and Business

Abstract Academia remains the central place of machine learning education. While academic culture is the predominant factor influencing the way we teach machine learning to students, many practitioners question this culture, claiming the lack of alignment between academic and business environments. Drawing on professional experiences from both sides of the chasm, we describe the main points of contention, in the hope that it will help better align academic syllabi with the expectations towards future machine learning practitioners. We also provide recommendations for teaching of the applied aspects of machine learning.