Three recent trends in the use of Technology in Healthcare

A little less than two months of 2020 have passed, and we already see some clear trends in the use of technology in healthcare. The healthcare industry has been one of the most benefited by the use of IT, which has helped in the improvement of diagnosis and personalized attention. Let’s see some of these trends.

Technology and home hospitalization

There’s an evolution in medical attention. Now, it’s not limited to physicians, nurses, and the hospital. Currently, homes are the perfect scenario to receive treatments and healthcare attention with comfort. The omnipresence of technology is leading to a transformation of the relationship between physician and patient, leading it to virtuality, which allows adapting the link to the characteristics of the patient, which brings greater comfort for the patient.

This may lead to hospitalization in a clinic, per se, to be reserved for complex situations of trauma, complicated surgeries, and emergencies, while other cares can be received from home.

Among the companies that have bet on this, we find HeOn – Health On Line, with HeOn Medical Home, which brings top-notch technological tools to transform patients’ homes into correct hospitalization environments.

Leveraging data

The healthcare industry won’t be left behind by the trend of data analytics. They are taking advantage of all of the technological and statistical resources to turn healthcare data in an infrastructure priority and a significant investment attractive.

This will allow patients, doctors, nurses, and other healthcare professionals to take advantage of the data to offer a new dimension to the diagnosis and the treatments towards the best possible option. We’ll also see the pharmaceutical industry taking advantage of this thanks to data management as it will allow the industry to work hand in hand with patients and healthcare providers to develop treatments in reasonable time and cost.

Use of wearables

Digital healthcare, as a concept, involves including IT in every step of the healthcare process (prevention, diagnosis, treatment, following, and management). One of these advances has been wearable technology. They began as small gadgets in your watch to know how many steps you took or how many calories did you burn when you went jogging. Healthcare is now applying this technology.

Some experts foresee that, for example, we’ll have skin patches manufactured with nanomaterials that will allow connection with IoT devices. For example, iRythm, a San Francisco-based company, launched its patch, Zio, which allows monitoring the cardiac rhythm for 14 days remotely. There’s been a radical cost reduction of these products. Hence, a market explosion could be right around the corner.

Conclusion

Healthcare has been one of the most benefited industries by the tech expansion of the last few years. However, a large part of its potential growth had slowed down because of the high costs of some technologies. Now, 2020 can be a vantage point because many technological applications have reduced their prices. We’re hoping that’s the case.

How will Bonjoro Leverage 5G for better
customer-oriented service?

When you receive a one-size-fits-all email from a company, you don’t open it. You just select the 3 or 4 emails you may have received and send them to the trash bin.

Would that be the case if, for example, you were interested in a Pilates membership, and you received a customized video addressing your questions about the membership? Probably not.

That’s what Bonjoro is aiming at. Bonjoro is a company launched in 2017 that uses 5G connectivity to produce personalized videos for customers.
Why is that important? Because it shows you care. It’s a basic principle of a customer-oriented strategy that integrates marketing and service-providing.
What justifies Bonjoro’s existence? The fact that there is a breach in the market. The harsh reality that email business comms are not doing their job. They are impersonal, and we are in the era of personalization and customization, which leads people to ignore them quickly. According to some studies, the average open rate for business chain emails is 37 %. That means they are far away from achieving their goal.

That’s where Bonjoro gets in. With 5G technology, they can produce personalized videos for customers. The topic? Whatever you wish: questions about products, customer assistance, general information, discounts, etc.

However, how can you make that possible? It depends on basically four factors: speed of recording, uploading, and streaming the video and the location of the business and the customer.

Thus, Bonjoro works way better for small businesses as they have a smaller customer base and, hence, can handle it in a much more personalized manner. However, 4G connectivity may not be enough for doing the job. For example, slow response times may hinder some operations, and using GIFs as video preview may be somewhat difficult.

But Bonjoro seems to be looking ahead of its time. In fact, the idea of personalized video email responses is not new. The people behind Bonjoro were also behind Verbate, a market research initiative that uses mobile video. So, yes, Bonjoro’s initiative may be a bit hard today due to connectivity issues but looking ahead… It seems like the way to go.

We are in an era where customer-oriented services need personalization. The ways companies communicate with their customers through email haven’t changed significantly in this century. And the rise of 5G may pave the way for these kinds of improvements.

Initiatives like Bonjoro help small businesses to balance the field between smaller businesses and large corporations. Yes, large brands sell status, but smaller businesses can sell personalization, which also causes a significant impact. And that personalization can come in many different ways: personalized email responses, ultra-fast customer service, etc.

No matter how it goes, one thing is for certain: The 5G era will transform the way businesses and customers relate to each other, enabling more personal interactions and helping businesses to have much more one-on-one relationships with their customers. Bonjoro is just one of the first few steps toward this radical transformation.

Ahora la tecnología te ayudará a brindar servicio al cliente 100% personalizado

Sé honesto, cuando recibes un correo electrónico masivo ¿Lo abres de inmediato? ¿O eres de los que selecciona los 3 o 4 correos masivos de tu bandeja de entrada y los envía directamente a la papelera?

¿Crees que pasaría los mismo, si por ejemplo, has solicitado información acerca de unas clases de Pilates en las que estás interesado y recibes un video en el que te hablan directamente de los temas de tu interés y te responden de manera personalizada las preguntas que has hecho?

Ese es el objetivo principal de Bonjoro, una compañía que nació en 2017 y que utiliza la conectividad de la red 5G para producir videos personalizados para clientes… y ¿cuál es el gran valor agregado de esta plataforma? ¡Que le da valor a los clientes! Es una iniciativa que basa su metodología en un principio de estrategia orientada al cliente, integrando el mercadeo digital y la prestación de servicios.

¿Qué justifica la existencia de Bonjoro? El hecho de que hay una brecha en el mercado. La dura realidad de las comunicaciones comerciales es que no están siendo las más efectivas debido a que son impersonales y estamos en la era de la personalización, esto lleva a los consumidores a ignorar muchos de los contenidos de sus correos electrónicos y descartarlos casi que automáticamente. Según algunos estudios, la tasa de apertura promedio para los correos electrónicos de la cadena empresarial es del 37%. Eso significa que están lejos de lograr su objetivo.

Es ahí donde entra Bonjoro a conquistar a los clientes. Con la tecnología 5G es posible producir con rapidez y precisión videos personalizados para los clientes abarcando cualquier tema que la marca considere relevante para su consumidor… preguntas sobre productos, asistencia al cliente, información general, descuentos, etc.

Sin embargo, ¿cómo es posible alcanzar el éxito? Depende básicamente de cuatro factores: velocidad de grabación, carga y transmisión del video y la ubicación de la empresa y el cliente.

Esto quiere decir que la plataforma funciona mejora para las pequeñas empresas, esto debido a que tienen una base de clientes más pequeña y, por lo tanto, pueden manejar sus comunicaciones de una manera mucho más personalizada. Sin embargo, la conectividad 4G puede no ser suficiente para hacer el trabajo. Por ejemplo, los tiempos de respuesta lentos pueden dificultar algunas operaciones, y usar GIF como vista previa de video puede ser algo difícil.

Pero Bonjoro parece estar anticipando su tiempo. De hecho, la idea de respuestas de correo electrónico de video personalizadas no es nueva. Las personas detrás de Bonjoro también estaban detrás de Verbate, una iniciativa de investigación de mercado que usa video móvil. Entonces, sí, la iniciativa de Bonjoro puede ser un poco difícil hoy debido a problemas de conectividad, pero mirando hacia el futuro, parece una solución bastante acertada.

Estamos en una era donde los servicios orientados al cliente necesitan personalización. Las formas en que las empresas se comunican con sus clientes a través del correo electrónico no han cambiado significativamente en este siglo. Y el aumento de 5G puede allanar el camino para este tipo de mejoras.

Iniciativas como Bonjoro ayudan a las pequeñas empresas a equilibrar el campo entre ellas y las grandes corporaciones. Si las mega-marcas pueden vender de manera masiva, las pequeñas organizaciones pueden estar orientadas a servicios personalizados, que podrían traducirse en una mayor fidelización de los clientes y relaciones más sostenibles. Sin importar cuál sea la forma en que las empresas y los clientes establecen sus relaciones, lo cierto es que la era 5G revolucionará la forma de interactuar entre marcas y mercado y Bonjoro es solo uno de los primeros pasos hacia esta transformación radical.

Could primary care improve the
United States’ Healthcare System?

If you take a good look at healthcare publicity on the Internet, TV, billboards, and radio, you’ll notice that specialists are the ones that are featured the most. You’ll see advertisements on cancer centers, plastic surgeons, cardiologists, ophthalmologists… However, you’ll find out that there’s one element missing: Primary care.

You hardly see primary care billboards and commercials. Nevertheless, primary care is incredibly important for improving the healthcare system. Why? Because primary care can prevent emergency room visits and hospitalizations, which means a lesser total expenditure. Primary care providers allow better early detection and treatment of medical conditions and preventive care.

In most parts of the US, primary care spending doesn’t surpass 7 %, which is less than half than most Western European countries, where it’s usually around 12 % and 15 %.

That means that about 1 in every 4 Americans don’t have adequate access to primary care. And it’s getting worse. In 2015, there was an average of five primary care practitioners per 100k persons. By 2030 the shortage will rise to 150k. Also, considering that most physicians are concentrated in urban and suburban areas, the situation will be even worse in rural areas, where people are, on average, older, poorer, and sicker. The picture is bleak.

What is causing this issue? Three factors mainly: lack of residency slots; levels of student debt that leads professionals into seeking higher-grossing specialties; and the large salary gap between primary care and specialists.

And… that’s not it. There are significant administrative difficulties related to the requirements to sustain an electronic medical record-keeping: primary care doctors have to work two hours filling bureaucratic documents for every hour they work with their patients. Yes, you read that correctly.

What can be done for primary care physicians? It’s not easy. Some new payment model propositions promote value over volume but, in fact, may increase bureaucratic burdens and practicing costs for primary care professionals. Also, a successful transition into a value-based model requires an adequate infrastructure of data exchange and care coordination, which, usually, doesn’t exist.

However, that doesn’t mean everything is lost. But the first step is clear: it’s necessary to increase primary care reimbursements. The only way to improve access to primary care is by having more primary care practitioners, and the only way to have more primary care practitioners is reducing the pay gap between primary care and specialties.

There are other radical proposals. The Lown Institute, for example, proposed that the Federal Government could cover medical school costs for students who dedicate ten or more years of their careers to primary care. They also recommend changing the primary care model into one that doesn’t require insurers. Primary care physicians could charge a monthly fee to clients for their services that would include all primary care services, which is the case in many European countries.

The American healthcare system is a hot-button topic among public policy specialists. However, those specialized in healthcare, agree on one thing: Primary care is key to improving the system, and it’s not working as it is. A value-based approach that gives incentives to practitioners may be the way to go.

¿Podrían los médicos primarios mejorar el Sistema de Salud de los Estados Unidos?

Si te fijas en la publicidad de los servicios de salud en internet, televisión, vallas publicitarias y radio, notarás que los especialistas ocupan casi todo el espacio. Es posible que veas anuncios de de centros oncológicos, cirujanos plásticos, cardiólogos, oftalmólogos, etc… Sin embargo, te darás cuenta que hay un elemento que se ve poco: Cuidados primarios.
A pesar de esta realidad publicitaria, es importante reconocer que los cuidados primarios son muy importantes para mejorar el sistema de salud ¿Por qué? Porque es en esta etapa de la prestación de los servicios médicos cuando se puede prevenir futuras visitas a la sala de emergencias y hospitalizaciones. Los proveedores de cuidados primarios ayudan a detectar y tratar condiciones médicas de manera temprana.

No obstante, en muchos estados de los Estados Unidos, el gasto en cuidados primarios no pasa del 7 % del gasto médico total. Esto significa que es menos de la mitad que en la mayoría de los países de Europa occidental, donde generalmente se encuentra entre 12 y 15 por ciento.

Dicho de otra manera, uno de cada cuatro estadounidenses no tiene acceso adecuado a los cuidados primarios y la situación está empeorando. En 2015 había cinco especialistas en cuidados primarios por cada 100 mil habitantes, para el 2030, la cifra aumentará a 150 mil. Considerando que la mayoría de los médicos están concentrados en zonas urbanas y suburbanas, podríamos asegurar que la situación es aún menos favorable en zonas rurales donde las personas son, en promedio, de mayor edad, menores ingresos y por lo general requieren en mayor medida de los servicios de salud. No hay buen pronóstico.

¿Qué está causando este problema? Fundamentalmente tres factores: pocas plazas para residentes; unos niveles de deuda estudiantil que llevan a los profesionales a buscar especializaciones que generen más ingresos; y la gran diferencia de salarios entre los cuidados primarios y los especialistas.

Por otra parte, también es importante tener en cuenta una serie de dificultades administrativas que obligan a los médicos primarios a pasar en promedio dos horas llenando documentos por cada hora que pasan con un paciente.

¿Qué puede hacerse para mejorar la situación? No es fácil. Algunas propuestas de pago promueven el valor sobre el volumen pero, de hecho, podrían incrementar las dificultades burocráticas para los médicos. También, una transición exitosa hacia un modelo basado en el valor sobre el volumen exige una infraestructura de intercambio de datos que, regularmente, no existe.
Sin embargo, eso no significa que todo esté perdido. Pero el primer paso es claro: es necesario invertir más en cuidados primarios. La única forma de mejorar el acceso a cuidados primarios es teniendo más profesionales de cuidados primarios, y la única forma de tener más profesionales de cuidados primarios es reduciendo la brecha de ingresos entre cuidados primarios y las especialidades.

Hay otras propuestas más radicales. El instituto Lown, por ejemplo, propuso que el gobierno federal cubriera los gastos médicos de estudiantes que dediquen diez o más años de su carrera a los cuidados primarios. También recomiendan pasar a un modelo de cuidados primarios que no requiera empresas de seguros. Las empresas de cuidados primarios podrían cobrar una mensualidad a los clientes por sus servicios. Este es el esquema que se sigue en la mayoría de los países de Europa.

El sistema norteamericano de salud es un tema prioritario para los especialistas en políticas públicas. Sin embargo, aquellos especializados en salud están de acuerdo en un aspecto: los cuidados primarios son fundamentales para mejorar el sistema y no están funcionando bajo el esquema actual. Un enfoque basado en el valor sobre el volumen que dé incentivos a los médicos para que participen en el sistema podría ser el camino a seguir.

Understanding the differences between Deep Learning and Machine Learning

AI advances constantly and it’s overwhelming to stay up to date. However, if you want to keep things simple you can reduce AI innovations in pretty much two notions: machine learning and deep learning. Aren’t they the same? Sometimes we have that impression but that’s a mistake, and is vital to know the differences.

Understanding the differences between them can help you understand how Facebook recognizes faces in pictures, what drives self-driving cars (pun intended) and how Netflix determines which shows you may like. Here we’ll let you know the main differences between machine learning and deep learning.

First things first, what are they?

Actually, deep learning is a particular type of machine learning. Both are forms of AI. That means that all deep learning is machine learning but not all machine learning is deep learning. They both begin with training, test data and a model. They both pass through an optimization procedure to lead the model to fit the data. They both handle regression and classification problems.

Now that we went through their fundamental similarities, we can explain better what they are. Machine learning algorithms are divided between supervised and unsupervised. Supervised problems are also divided into classification and regression problems. Classification problems predict non-numeric answers and regression predict numeric answers.

Unsupervised learning is divided into clustering (identifying groups of similar objects), association (identifying sequences of objects), and dimensionality reduction (such as projection, and characteristics selection and extraction).

On the other hand, deep learning is a particular form of machine learning where the model has more than one hidden layer between the problem and the solution. That means that a deep learning model algorithm can determine if a prediction if accurate or not, which means it’s more advanced that basic machine learning algorithms. They are designed in such a way they analyze data through a logical structure that resembles the one of a human being.

An awesome example of deep learning is AlphaGo, from Google. The company created a program that had an engrained neural network that learned to play Go, the board game, which requires deep intuition. AlphaGo played against pro players, and the deep learning algorithm played at an unparallel level without being told how and when to make a specific move.

The differences

First of all, as we said before, deep learning is a more advanced subset of machine learning. Typical machine learning uses algorithms to analyze the data, learn from it, and make decisions based on what it learns. Deep learning, instead, organizes algorithms in different layers, a structure called “neural network” that learns and makes decisions independently. That means that even as all machine learning is AI, deep learning is the one that sustains almost all human-like AI.

That means deep learning can discover features that it needs to classify, while machine learning needs to be provided with these elements manually. In consequence, deep learning requires high-end technology and large amounts of data to deliver exact, human-resembling, results.

ClassPass becomes the newest unicorn: How did they get there?

ClassPass became the first unicorn of the decade last week when it passed the billion-dollar mark after announcing a $285 million Series E investment that put their valuation over the billion-dollar threshold. Also, it’s quite fitting that the founder of the first tech unicorn of the 2020s is a woman of color.

“We are motivated by the impact we’ve had on members and partners, including 100 million hours of workouts that have already been booked,” said founder Payal Kadakia in a statement about the latest investment. “This investment is a significant milestone that will further our mission to help people stay active and spend their time meaningfully.”

How did ClassPass become a unicorn in a little more than six years? Here we’ll explain what leadership decisions got the company to this point.

First of all,  what’s ClassPass?

ClasPass is an American fitness company that provides access to many different fitness and wellness classes such as yoga, martial arts, boxing, pilates, and the use of health clubs through a flat-rate monthly subscription.

It was founded by Payal Kadaka and Mary Biggins and is the globe’s largest wellness club aggregator, and it expanded from 4 countries to 28 in the last 18 months. At the same time, they’ve entered the corporate wellness business partnering with 1000 companies such as Google and Morgan Stanley.

How did they get there?

First of all, they identified a breach in the fitness industry: there was no aggregation. Consumers wanted variety, and they weren’t able to get it in one single place. ClassPass gave them a chance to pick from a vast catalog of fitness and wellness activity so that they could personalize their training schedule for a flat-rate subscription.

After initial success, they were bold. Growing from 4 countries to 28 was a significant bet, to say the least. But when you’re a business leader, you’ve got to trust your product. That helped them launch SmartTools, which helped their more than 30,000 business partners (gyms, health centers, spas) to increase their revenues with better yield optimization. That, in return, will allow ClassPass to improve their variety of offerings, and give a better experience to their subscribers.

The Key: Service-Focused Strategy

Payal Kadakia, the founder of the company, has always said ClassPass is a mission-oriented company with two primary purposes: 1) helping people around the world to get fitness and 2) helping small fitness and wellness businesses grow. It’s pretty simple: when you offer an excellent service and support your partners, so they provide a great service, everyone is happy, everyone is making money, and everyone is satisfied. The subscribers get better, varied, and more personalized business experiences, partners thrive, and ClassPass grows. A killer combination.

What can we learn from this?

The ClassPass case gives two great lessons for every organizational leader. First, trust the quality of your product and don’t be afraid to be bold in your decisions, and, second, use a service-driven strategy. Always seek to offer the best product and the best experience and focus on helping your partners grow, too, as it will benefit your clients, too.

Sustainability, a strategy to have a profitable business

The Veja sneaker brand teaches that it can be scaled in the market with sustainable production and without investing in advertising thanks to the new “green” trends.

Many of the most famous and globally positioned companies are changing their products towards more sustainable ones, however, many of these strategies remain only in environmentally friendly packaging. Although these types of changes are necessary and appreciated, there are brands that are going further, since they understand that the consumer seeks a comprehensive change in the way in which the product they are buying is made.

From chocolates that do not destroy any species or ecosystems in their production, through eggs from chickens that are not tortured, to slippers whose manufacture is not linked to the exploitation of the environment or the workers who are responsible for the production process, such as Veja case. These shoes, rather than being made exclusively of environmentally friendly materials, every step of the production process is meticulously cared for in a way that is entirely sustainable, ethical and human.

The name Veja comes from Portuguese and means “Mira”, where the founders wanted to express a literal message “Look how your shoes are made”. And this claim can fortunately be allowed with property since its philosophy is leveraged in fair trade where they seek that from the worker who is responsible for producing organic cotton, through supply and packaging, to distributors obtain monetary rewards human, fair and proportional to the work done by them. At the same time, it takes into account details such as space resources, which is why they work with a green electricity cooperative (ENERCOOP), as well as are allies of “Ateliers Sans Frontières” that is presented as an organization that is responsible for helping and reintroduce people with legal backgrounds or vices such as drugs or alcohol to the labor market and, consequently, to society.

This is where the most frequent question comes … If everything is so perfect, it should cost a fortune, isn’t it? The answer is yes. The costs of these shoes are about four times higher than

Some that are not sustainable. However, to keep prices accessible to the masses, they use as a strategy not to invest a penny in advertising. It seems like it doesn’t make much sense, right? The truth is that since the first quarter of 2019 they have positioned themselves as the most desired shoes by bloggers and Instagram personalities.

Regarding this strategy, its co-founder, Sébastien Kopp, stressed that it is not worth investing in advertising if those costs would be even higher than those of shoe production. Thus they concluded that, similarly, word of mouth is the most powerful form of advertising, and there they were right because, thanks to their ecofriendly practices, consumers have made known their different products and, right now, they are fever Instagram

Sustainability is the future and Veja teaches us in the best way that a brand can be socially and environmentally responsible while remaining profitable and recognized worldwide.

¿Cómo migrar a la nube y no morir en el intento?

Hasta no hace mucho, usar la nube era una opción secundaria en la industria tecnológica. Sin embargo, en el último par de años, la tecnología de nube se convirtió en la primera opción para muchas compañías, tomando en cuenta la flexiibilidad, relación precio-valor y escalabilidad que provee.

Todo eso significa que, prácticamente, cualquier compañía migrará a la nube más temprano que tarde. Sin embargo, la migración hacia la nube no es rápida o sencilla. Errores en el proceso pueden poner en peligro toda la infraestructura tecnológica de una organización. Eso no debería detener a nadie en migrar a la nube, pero significa que es necesario tomar un enfoque racional y bien planificado frente al proceso. Acá discutiremos algunas cosas al respecto.

Algunos retos a considerar

La tecnología de nube puede reemplazar a los servidores on-site pero también genera un enfoque completamente distinto respecto al desarrollo de la tecnología en la compañía. Esta es la distinción más básica para entender la migración a la nube. No es simplemente mover servidores e información de un lugar a otro. Es un proceso holístico de simplificación y mejora para reducir costos, agilizar procesos y potenciar las capacidades de tu negocio.

Por este motivo suelen llamar a la migración a la nube “transformación digital”. Usar la nube es mucho más que cambiar tu base de datos; es, más bien, una nueva perspectiva que afecta cada parte de tu infraestructura tecnológica. Por esto es que adquirir una perspectiva holística es uno de los grandes retos que tendrás.

¿Cómo migrar a la nube correctamente?

Para conducir una migración a la nube exitosa, debes seguir unos principios básicos.

Primero, haz un inventario completo de todos tus recursos. Si excluyes alguno, estás comprometiendo toda tu infraestructura tecnológica y tu migración. Esa es la razón principal por la cual debes hacer un inventario completo de todo lo que necesitas para hacer la transición.

Luego, debes saber que necesitas tomar un enfoque holístico. No es sobre mover de una forma de almacenamiento a otra. No. Es sobre mejorar procesos. Por ejemplo, puedes reducir el número de roles de servidores si usas este enfoque adecuadamente.

También, considera todas las opciones. Hay muchos caminos que puedes escoger. De nuevo, toma un enfoque holístico. No lo veas como un proceso estándar, sino que busca cada oportunidad que puedas tener para adquirir nuevas capacidades y adoptar mejores prácticas. No te enfoques exclusivamente en reducir costos si no en cómo la migración puede ayudarte a mejorar tu estructura tecnológica.

Por último, también es esencial crear una estrategia de seguridad. La ciberseguridad en la nube es sumamente diferente de la ciberseguridad on-premise. Las filtraciones de datos son comunes, pero no es culpa de la nube. Esto ocurre fundamentalmente por falta de organización al migrar, lo que deja varias partes de la infraestructura vulnerable ante ataques.

En la nube, los recursos no se encuentran protegidos por un firewall, por lo cual, es necesario utilizar un firewall de nube para protegerlos de filtraciones.

En conclusión

Migrar a la nube puede no ser el proceso más sencillo, pero hoy día, es indispensable. Sus beneficios son mucho mayores a sus riesgos si tomas un enfoque holístico, bien planificado ante pasar de una tecnología on premise a una basada en la nube. Te permitirá reducir costos, mejorar la escalabilidad y flexibilidad y obtener un mejor desempeño.

Conoce las diferencias entre deep
learning y machine learning

La inteligencia artificial avanza constantemente y es abrumador mantenerse actualizado. Sin embargo, si deseas simplificar los conceptos de inteligencia artificial, puedes reducirlo todo en dos principales: machine learning y deep learning. ¿Acaso no son lo mismo? A veces pensamos eso, pero es un error y es fundamental saber las diferencias.

Comprender las diferencias entre ellos puede ayudarte a comprender cómo Facebook reconoce rostros en fotos, cómo funcionan los vehículos autónomos y cómo Netflix determina qué programas podrían gustarte. Acá te explicaremos las principales diferencias entre machine learning y deep learning.

Primero que todo, ¿qué son?

En realidad, el deep learning es solo una forma de machine learning y ambos son formas de inteligencia artificial. Eso significa que todo deep learning es machine learning pero no todo machine learning es deep learning. Ambos comienzan con entrenamiento, prueba de datos y un modelo. Ambos pasan por un proceso de optimización para que el modelo se adecúe a la data. Ambos manejan problemas de regresión y clasificación.

Ahora que ya explicamos sus similitudes fundamentales, podremos explicar mejor qué son. Los algoritmos de machine learning se dividen en supervisados y no-supervisados. Los problemas supervisados se dividen en problemas de clasificación, que predicen respuestas no-numéricas y de regresión, que predicen respuestas numéricas.

Los algoritmos no-supervisados se dividen en agrupamiento (identificar grupos de objetos similares), asociación (identificar secuencias de objetos) y reducción de dimensiones (como proyección y selección y extracción de características).

Por otra parte, deep learning es una forma particular de machine learning donde el modelo tiene más de una capa entre el problema y la solución. Esto significa que un algoritmo de un modelo de deep learning puede determinar si una predicción es acertada o no, lo cual significa que es más avanzado que los algoritmos de machine learning básicos. Están diseñados de forma que analizan la data a través de una estructura lógica similar a la de un ser humano.

Un maravilloso ejemplo de deep learning es AlphaGo, de Google. La compañía creó un programa que incluía una red neuronal y aprendió a jugar el juego de mesa Go, que requiere profundas habilidades de intuición. AlphaGo jugó contra jugadores profesionales y el algoritmo llegó a jugar a un nivel sin comparación sin que le dijesen cuándo y cómo hacer una movida específica.

Las diferencias

Primero que todo, como dijimos anteriormente, el deep learning es una forma más avanzada de machine learning. El machine learning típico usa algoritmos para analizar la data, aprender de ella y tomar decisiones basadas en lo que aprende. El deep learning, en cambio, organiza algoritmos en distintas capas. Esta estructura se llama “red neuronal” que aprende y toma decisiones de forma autónoma. Eso significa que, aunque todo machine learning es inteligencia artificial, el deep learning es la base de casi toda la inteligencia artificial similar a la lógica humana.

Eso significa que el deep learning puede descubrir elementos que clasificar, mientras que el machine learning requiere que dichos elementos se le provean manualmente. En consecuencia, el deep learning requiere maquinaria de punta y grandes cantidades de datos para entregar resultados que se asemejen a la lógica humana.