In natural language processing (NLP), word embedding is a term used for the representation of words for text analysis, typically in the form of a real-valued vector that encodes the meaning of the word such that the words that are closer in the vector space are expected to be similar in meaning. Thus word embedding is the process of converting a word into a computer-comprehensible expression before the next operation can be performed.

https://machinelearningmastery.com/what-are-word-embeddings/

https://zhuanlan.zhihu.com/p/27830489

Bayes' theorem, named after Thomas Bayes, describes the probability of an event, based on prior knowledge of conditions that might be related to the event.^[2] For example, if the risk of developing health problems is known to increase with age, Bayes' theorem allows the risk to an individual of a known age to be assessed more accurately (by conditioning it on their age) than simply assuming that the individual is typical of the population as a whole.

DeepMind, a subsidiary of Google, is a company that creates computer algorithms that use neural networks and machine learning to play games like chess, Go, Shogi, and other strategic board games. 

Here is a link to an article that explains AlphaZero, the chess engine developed by DeepMind:

https://www.chessprogramming.org/AlphaZero#Network_Architecture

Microcontrollers are small, low power, inexpensive processors that control basic functions, for example the computer that controls a microwave oven. These processors are incredibly inexpensive (powerful ones cost under $10), but don't have nearly the power of a processor that would be found in a phone or laptop. 

Machine learning models can be lightweight ways to implement complex computations that the processors would otherwise be incapable of. The use of machine learning models on small, local devices without unnecessary internet connections or overpowered hardware can increase privacy and efficiency of many household products.

https://www.tensorflow.org/lite/microcontrollers

FINAL TOPIC: I think I will be learning about demixing and some examples of where it has been/can be used (such as in The Beatles: Get Back documentary, which I did some research into last week).

FINAL TOPIC: I plan to research how machine learning is used in the speech recognition utilized by virtual assistants like Google's Assistant, Apple's Siri, and Amazon's Alexa. Machine learning is used prominently in helping these assistants recognize and interpret commands from the user, including a variety of different phrases and instructions.

Some sources: 

https://towardsdatascience.com/conversational-ai-intelligent-virtual-assistants-and-the-road-ahead-6345db47d106

https://www.ttec.com/glossary/intelligent-virtual-assistant

https://machinelearning.apple.com/research/hey-siri

"Hey Siri, how's the weather?" If we're iPhone users, we've likely tried Apple's own virtual assistant, Siri. In fact, we can use Siri for almost anything, such as finding GPS navigation, making a call or sending a text message, and many other functions that are made easier thanks to Siri's voice recognition software. However, how does Siri actually work?

The algorithm of Siri:

https://medium.com/mytake/natural-language-processing-ft-siri-2bc7b854a2a3

How does Siri work in NLP:

https://www.ibm.com/cloud/learn/natural-language-processing

https://skywell.software/blog/how-does-siri-work-technology-and-algorithm/

Voice Trigger about "hey Siri" in personal assistant:

  https://machinelearning.apple.com/research/hey-siri

Along with the development of civil aircrafts, machine learning helps the crew members to do a lot. 

For example, automated system have been an area of concern in commercial aviation for years, with the adoption of fly-by-wire control and automated flight system, machine learning and artificial intelligence technologies are taking on the workload of the crew members in the cockpit, and even becoming another co-pilot. 

Articles:

https://www.researchgate.net/publication/262764214_Machine_Learning_Model_for_Aircraft_Performances

https://www.altexsoft.com/blog/engineering/ai-airlines/

Machine learning is beginning to be used in the games industry both as part of games and as part of the development process for games.

1. For example, a game currently in development called Hello Neighbor 2 is using machine learning to create an opponent for the player that learns from the actions of all of the game's players.

https://screenrant.com/hello-neighbor-2-gameplay-ai-trailer-mr-peterson/

https://www.youtube.com/watch?v=Hlj2SyGnlvc

2. Google has also found that neural networks can be used in place of human playtesters.

https://ai.googleblog.com/2021/03/leveraging-machine-learning-for-game.html

3. Below are some articles further explaining examples and challenges with using machine learning in video games.

https://www.theverge.com/2019/3/6/18222203/video-game-ai-future-procedural-generation-deep-learning

https://www.caltech.edu/about/news/neural-networks-playing-video-games-teach-us-about-our-own-brains

https://futurism.com/the-byte/browser-game-opponents-neural-networks

Peter Jackson, the director of the 2021 documentary series The Beatles: Get Back, developed a machine learning system to separate the sound of The Beatles' voices from the sound of their instruments (a process known as demixing) to clean up 150 hours of rehearsal audio recorded in mono in 1969.

https://www.slashfilm.com/677857/the-beatles-get-back-used-high-tech-machine-learning-to-restore-the-audio/

https://www.morson.com/blog/2021/12/machine-learning-artificial-intelligence-films-Beatles

https://collider.com/peter-jackson-the-beatles-get-back-interview-demixing/

https://musictech.com/news/music/peter-jacksons-beatles-documentary-uses-ai-to-reveal-hidden-studio-conversations/

Since it first launched in 2016, TikTok has amassed more than 1 billion monthly users. The key to its success lies in its addictively scrollable and endless stream of videos seemingly targeted right at all different kind of people. 

the links below are some basic information about how the recommendation algorithm used by Tik Tok affected people's behaviors.

why it is so addicted:

Why TikTok's algorithm is so addictive | Popular Science (popsci.com)

collected data:

How Does The TikTok Algorithm Work? Here's What You Need To Know (later.com)

Compared Tik Tok and Youtube.

Americans are spending more time on TikTok than YouTube, report says (inputmag.com)

Waymo LLC is a company owned by Google that is using machine learning to create completely autonomous vehicles. They currently offer driverless taxi services in certain cites in the US.

https://waymo.com/waymo-driver/

https://www.futurecar.com/3929/Waymo-Develops-a-Machine-Learning-Model-to-Predict-the-Behavior-of-Other-Road-Users-for-its-Self-Driving-Vehicles#:~:text=com%20%2D%20via%20%40FutureCar_Media-,Waymo%20Develops%20a%20Machine%20Learning%20Model%20to%20Predict%20the%20Behavior,for%20its%20Self%2DDriving%20Vehicles&text=By%20anticipating%20this%20behavior%20the,working%20on%20self%2Ddriving%20vehicles.

https://www.analyticssteps.com/blogs/how-waymo-using-ai-autonomous-driving

I currently have a job with Collins Aerospace in Cedar Rapids where we work in SAP and other systems to input data and process transactions for the company. I work on analyzing different supply chain metrics using data extracted from SAP to find different improvements that can be made. For this reason, I am interested in some use cases of machine learning in SAP.

1. Data Attribute Recommendation

There are two main types of data companies use: transactional data and master data. Transactional data are those tied to different transactions (invoices, sales orders, movement of product, etc.). Master data often is used to describe parts, customers, suppliers, etc. Master data does not change often as opposed to transactional data. Some of the main issues are the large maintenance and the inconsistency of master data which can cause delays to processing orders or shipments. Data Attribute Recommendation uses free text, numbers and categories as input to classify entities such as products, stores and users into multiple classes and also to predict the value of missing numerical attributes in your data records. You can use Data Attribute Recommendation, for example, to classify incoming product information and predict the price of commodities based on their description. 

https://discovery-center.cloud.sap/serviceCatalog/data-attribute-recommendation/?region=all&tab=feature

2. Purchase Order Processing

For the procurement organizations of manufacturing companies, the timely availability of purchased goods and reliable planning are key for smooth production processes. Operational purchasers need to prevent production delays and rescheduling. Using machine learning technologies to automatically predict delays in supplier deliveries, manufacturers can prevent gaps in raw material availability and avoid production rescheduling.

https://www.sap.com/dmc/exp/2021-09-75642-intelligent-all-use-cases/index.html#purchase_order_processing

Google Maps uses machine learning algorithms to predict frequent locations and ETA times for users of the service. These articles go in-depth on how Google Maps uses machine learning to improve its efficiency. 

https://www.springboard.com/blog/ai-machine-learning/machine-learning-google-maps/#:~:text=Google%20Maps%20will%20has%20predictive,in%20hundreds%20of%20cities%20worldwide.

https://www.theverge.com/2020/9/3/21419632/how-google-maps-predicts-traffic-eta-ai-machine-learning-deepmind

https://www.technologyreview.com/2021/06/23/1026653/using-machine-learning-to-build-maps-that-give-smarter-driving-advice/

You could begin your search for a topic by identifying employers that look attractive to you, and then looking for applications of machine learning at those companies:

• The Best Software Companies to Work for in 2020

Or maybe you will begin with an application of machine learning that you find especially interesting, and then look for companies or universities that are exploring that kind of application:

• What is Natural Language Processing (NLP): what it is and why it matters

A search that includes words like "customer success story" might also be productive:

• Deep Learning Success Stories