Research Hub.

Human-computer interaction (HCI) is the study and design of how people interact with computers and other similar systems. Its main focus is creating interfaces that help people interact with technology effectively. Because HCI is related to humans and technology, it has several disciplines, like computer science, psychology, ergonomics, graphic and product design, anthropology, and engineering

Usability means how easy and pleasant to use a certain system, while utility means whether the system has all the features that a person needs. A useful system needs to have both. For example, a calculator app is usually simple and easy to use, making it usable. But wont be useful to someone who wants to write and test Python programs because a calculator does not provide the features needed

An example of a confusing interface is a microwave with so many buttons, symbols, and cooking modes. A person who is gonna use the microwave may only want to heat his food for 30 sec, but he needs to navigate through the options to find his option; this creates unnecessary complexity. A good interface design should make the important and mainly used action easy to find. One improvement would be to make the common controls, like start and stop, more noticeable while keeping less commonly used options separate.

A prototype is an early version of the interactive design that allows people to experience how a product works before developing the final product. Designers would give users tasks and observe where they hesitate, make mistakes, or even be confused. This allows problems to be found and fixed before launching the final product. Just asking Do you like it would not help in finding out if they can complete important tasks with the interface.

One example is speech recognition, which enables people to control a computer using voice commands. It can be used to type text, navigate through menus, activate buttons, and other different commands. According to W3C, this can significantly help people with physical disabilities who can't easily use a keyboard and mouse, people with repetitive stress injuries, and people who find speaking easier than typing. To test it usefulness, i would ask the users to complete the same tasks using their voice and then usijgn a keyboard and mouse. I would compare the time taken to complete each task, what errors occurred, and how easy the users find each method.

We can split the problem among many connected computers; each one of them handles a part of the problem, and they communicate with each other to combine their results. This is called distributed computing; it allows us to handle problems that require more processing or memory capability of a single computer. However, the problem needs to be able to be split up.

Not necessarily; there might be some parts of the task that need to be done in order, so not all of the computational power would be used simultaneously. Computers will also spend time communicating, sharing data, and waiting for each other; these delays would mean that 1000 computers working together wouldn't mean finishing 1000 times faster. The limit that is caused by sequential parts is called Amdahl’s law.

Yes, under certain conditions, they can use consensus algorithms, which set rules for reaching an agreement. Algorithms that handle computers that send wrong or conflicting information with other computers provide Byzantine fault tolerance. For example, in a classical model with good communication and no unforgeable signatures, 1000 computers can handle up to 333 faulty computers. The guarantee depends on the algorithms and its assumptions about communication and failures

The computation is shared between my device and a remote server; my device runs the interface and my input, while the servers in the data centers do most of the processing and storage. These servers can work together from different locations. In most of the online games, my device would handle drawing the graphics while the server manages the shared game state, like a player’s location.

One possibility is building a much more detailed virtual model of biological processes, which would help scientists investigate diseases and test potential medicines at a scale that is currently not possible. Distributed computing already helps in this: Folding@Home connects computers to simulate how a protein moves. So I believe that larger, more dependable systems can extend this work to more complex biological models, but having greater computing power alone would not mean reaching medical breakthroughs

A modality in AI is a type of information that an AI system can process and understand. Different modalities represent information in different ways. Some examples include text such as messages and documents, images like photographs and diagrams, and audio like speech and music.

Multimodal learning is when an AI system learns from and combines different types of information. For example, an AI system can process an image that includes a written description of the photo; by combining the different information, the system can better understand what it is analyzing.

Multimodal learning is used in healthcare, robotics, virtual assistants, and self-driving cars. A real example of this is the Waymo Driver, which is used in Waymo’s self-driving cars. The system combines information from the camera, LiDAR, and radar. The camera provides visual information of the road traffic light, pedestrians, and other cars, while LiDAR provides the 3d information regarding the shape and distance of the cars or pedestrians; radar helps in measuring the speed and distance of the surrounding objects. Combining these different sources of information helps the car understand its surroundings.

Embodied AI is AI that can interact with its surroundings through a physical or virtual body. Instead of just processing information, embodied AI can look at its surroundings, make decisions, and perform tasks. For example, the AI could look at an object, decide how it would reach out to it, and then move its robotic arm to grab it.

Multimodal learning and embodied AI are connected because embodied AI often needs to process different types of information it understands and be able to interact with its surroundings. For example, a robot might use its camera to see objects, a microphone to hear instructions, language to understand commands, and physical sensors to know what is around it. Multimodal learning helps the robot to combine this information, while embodied AI uses it own understanding to decide what it needs to do next in its environment.

Artificial intelligence (AI) is the technology that allows computers and machines to do tasks that would normally need human intelligence; this can include learning from information, finding patterns, understanding language, and solving problems. AI doesn't necessarily mean that the machine thinks like a human. But means that the machine can do certain tasks that require a type of intelligence

The three major subfields of AI are machine learning, computer vision, and natural language processing. Machine learning means that computers can improve from the data that they were trained on without being programmed for this situation. Computer vision means that machines can understand information that comes in the form of images and videos by recognizing the person or object. Natural language processing, or NLP, helps computers understand and interpret human language,m like in chatbots and voice assistants

AI is booming right now because many of the technologies have improved at the same time; computers and specific hardware have become powerful, allowing researchers to train much larger AI models. Also, there is a huge amount of digital data that is available to train AI models. At the same time, advances in machine learning algorithms, mainly deep learning, have significantly improved what AI can do. Cloud computing helped in making powerful computing resources easier to access. Finally, large investments from tech companies and governments helped accelerate AI research and development.

Robots are used in manufacturing, logistics, and healthcare. In manufacturing, robots are used for tasks such as welding, assembling products, and handling materials because they can do repetitive tasks quickly, accurately, and consistently. In logistics, robots are used to transport goods, arrange products, and assist in deliveries because they can reduce repetitive physical work, which improves efficiency. In healthcare, robots can help with surgery, rehabilitation, and lab work. Robots are useful in these areas because they can do the repetitive, precise, and potentially dangerous tasks while allowing humans to focus on more complicated work.

The first major challenge is navigation and localization. In large places like the Mall of Qatar, the robot needs to know where it is and find out how to reach the different areas without getting lost. The second major challenge is avoid people and the other obstacle, the mall is a changing enovrioemt where people, carts and cleaning equipment constanlky move, so the robot must detect them and safelky change it path to avoid them. The third major challenge is battery life and continuous operation; a robot would not be able to operate for 24 hours normally on one battery charge. So it needs to monitor its own battery, then go to the nearest charging station if it needs to, and then continue its surveillance task. The robot would need reliable sensors and software because a failure during a 24h surveillance task would hinder it from completing the task successfully.

A decision problem is a problem that can be answered with only a yes or no. For example, “Is 17 a prime number?” This question can be answered only with a yes-or-no response. In computer science, most of the problems are written in this format because it is easier to study how difficult they are for computers to solve.

A problem can be considered decibel if their is an algorithm that cna always output the correst yes or no answer and hten stop for every possible input. No t every pr0blem is considered decidable. For example, the Halting problem cannot be solved by one algorithm for every possible program and input.

P is the class of decision problems that can be solved by computers in polynomial time, implying that the time needed grows at a manageable rate as the size of the input increases. NP is the class of decision problems whose possible solutions can be checked in polynomial time when given to a computer. All problems in the class P are also in NP because if you can solve something quickly, you can also check its solution quickly.

The idea behind P vs NP is that if a solution can be checked quickly can it also be found quickly? We understand that P is inside of NP but we dont know if P and NP are equal, If P=NP, many problems that look very difficult can be solved efficiently, as of today, the problem is not solved.

If i was able to solve the problem and my solution was accepted according ot the Clay Mathematics Institute's rules, I would receive 1 million dollars. P vs NP is considered one of hte seven Millennium Prize Problems, and 1 million dollars was put under each problem. So other than being known in the computer science community i would recive 1 millon dollars.

Because punch cards were slow, inconvenient, and very difficult to fix when a mistake happened. Programming languages made it easier to write instructions to the computer in a way humans can also understand. Then, the computer translates the instructions we wrote into machine code. This shows that the purpose of programming languages is to ease communication between computers and humans.

Programming languages are designed for different purposes; some are designed for efficiency, some for readability, and others for specific things like creating websites, databases, or operating systems. For example, Python is designed to be simple and readable, while C gives the programmer more control over the computer; different programming languages let programmers choose the one that fits their needs.

I only use Python, and one drawback is that data-type bugs can only be found when I run the program. For example, adding “5” and 5 looks the same, but one is an integer, and the other is a string. Python is also considerably slower than C or C++; it would be better if Python had a stricter mode that could detect bugs before the program ran.

First, I would ask myself why I am creating a new language and what problem I am trying to solve; then I would start to define its syntax, data types, variables, functions, conditions, and loops, and decide between it being compiled or interpreted and how to deal with errors, After that, I would create a simple interpreter or compiler and slowly add more features as the language develops.