Flashcards
Sem título
Flashcards do conjunto (26)
- 1uninformed/blind searchDoes not contain any domain knowledge. It operates in a brute force. It examines each node of the tree until it achieves
the goal node. - 2what type of uninformed search are there?•Breadth-first search: explores all the neighbour nodes in the current depth level before progressing to the nodes
existing in the next depth level.
- used for navigation system
•Depth-first search: goes as deep as possible down one path before backing up and trying a different one.
-used for solving maze - 3list advantages and disadvantages of blind search• Guarantee to find a solution but very slow.
• Time complexity is very high.
• Memory usage is very high (computer can’t execute the algorithm).
• Several problems are very hard to solve exactly. - 4what is heuristic/informed searchuses domain knowledge (use extra information
to guide the search). - 5what types of heuristic search techniques are there?➢ Local search techniques: Consider one solution at a time: only moving to the next solution if its better than the current one
➢ Population-based techniques: Consider more than one solution at a time. Inspired by human evolution. - 6Give me a example of local search techniques, name its disadvantags and how to cope with theseHill Climbing. can be stuck in local optima
cope with this limitation by e.g. decrease in value(going down hill) - 7Give me example of population based techniqueGenetic algorithm:
1. Create (randomly) a set of initial solutions as the first generation.
2. Perform crossover and mutation between these solutions to generate new generations.
3. Keep doing this until you reach the best solution. A stopping condition needs to be defined. - 8whats the difference between mono and multi objective optimization➢ Mono-objective optimization
• There is only one objective function that needs to be optimized.
• The primary goal is to find the best solution that maximizes or minimizes the single objective function
➢ Multi-objective optimization
• There is more than one objective function to be optimized simultaneously.
• The goal is to find a set of solutions that represent a trade-off among the conflicting objectives, rather than a
single optimal solution. - 9Explain machine learningits a sub category of AI
uses data
taking over in terms of popularity. - 10what type of algorithms are there in supervised learning?K-nearest neighbour(K-NN): lazy learner. Now, given an unclassified point, we can assign it to a group by observing what group its nearest neighbors belong to. This means a point close to a cluster of points classified as ‘Red’ has a higher probability of getting classified as ‘Red’.
Decision tree: eager learner. A Decision Tree uses a set of if-else rules to classify data based on features. It splits the data recursively to create branches - 11What methods are there within machine learning?supervised:we give the computer a large amount of labelled data. we feed it something, then ask to memorize it and make predictions of it.
unsupervised: no labels, there is not always historical data, like in medical fields, we then feed the algorithm data and ask it to identify a pattern.
Reinforcement: Learn from mistakes. Put AI in enviornment, give it positive/negative rewards. - 12What methods are there within unsupervised learning?Partitional-based clustering technique: K-Means- groups the unlabeled dataset into different clusters based on the distance between their centroids. The centroid, or cluster center, is either the mean or median of all the points within the cluster depending on the characteristics of the data. sensitive to outliers (noise)
Density-based clustering technique--DBSCAN. good for clustering applications with noise and spherical shape. - 13What type of methods are there withing recommender systemApplication that provides personalised suggestions or
recommendations to users.
Predict what items a user might be interested in based
on their preferences, behaviour, or interactions with the system.
Collaborative filtering: user based (recommended product based on similar user) itembased (look at other items u bought before)
Content based filtering: dive into actual item, and look into the content and filter based on it
Both supervised and unsupervised - 14Explain ethics of AIOntop of accuracy we also expect them to be fair, privacy preserving, explainable (we want to understand it)- this is tumor, as a doctor i would like to know why the algorithm reached that decision. another layer of complexity
- 15The three question to ensure ethical principle is followed.-The company you're working for is of negative/positive value. Is there a positive use for it? eg. bombs are a negative value. Is it for good or not.
-Is it for good? Is there a possibility that the ai system can be unfair? e.g. requiring people, is there a possibility of unfairness?
-is there a possibility that it might be used for other uses besides the intended usage? in the future it might be misused in a group or by yourself? - 16What can developers do to reach trustworthy ai?minimize the data to a minimum. e.g. app, require the minimum from the user. e.g. age, is it really necessary? do they give u an option to not answer?
even if its needed, it needs to only be used for the duration of usage, when u delete it, the information should also be deleted
can i augment the data, instead of using the real data, can i use synthetic data?instead of using real data, can we replace it?
synthetic, compared to augmented, is completely made up from scratch. to use real data u need consent etc. we deal with preserving the privacy and fairness as well
cleaning, contribute to the fairness - 17Explain transparency within aiinterpretable ai: can i implement extra measures to understand the process
explainable ai; understand the actual predictions, is there a tumor or non tumor? the algorithm says this is a tumor and why that conclusion is reached. it took this feature, and reached this point bc of this.
bright color on tumor, which makes the doctor believe the answer more - 18Explain the difference between state space and search tree• The state space is the set of all valid states that can exist in our
“problem universe”.
• Each node in the state space corresponds to an individual state.
• Nodes are directionally connected according by state transforming operators.
• The search tree corresponds to the sequence of paths that can be
reached from a given start state.
• A search tree is rooted at a particular start state.
• Each node in a search tree corresponds to a sequence of states (path) from the
start state (root) to a given state. - 19What is Search algorithm?• A search algorithm is a type of algorithm used in artificial
intelligence to find the best or most optimal solution to a
problem by exploring a set of possible solutions, also
called a search space.
• A search algorithm filters through a large number of
possibilities to find a solution that works best for a given
set of constraints. - 20What types of costs need to be considered when applying search algorithm?- Search Cost – The cost of finding a solution
- Solution Costs – The cost of using this solution (how good is the
solution) - 21What type of evaluation is there in relation to search methods?• Time complexity: How long will it take to find a solution? (This is related to
search cost)
• Space complexity: How much memory will be used?
• Completeness: When success is possible, is it guaranteed?
• Optimality: Will the best (min solution cost) solution be found? (This will be
related to solution cost) - 22What advantages / disadvantages are there to BFS• Advantages:
• BFS will provide a solution if any solution exists
• If there are more than one solutions for a given problem, then
BFS will provide the minimal solution which requires the least
number of steps
• Disadvantages:
• It requires lots of memory since each level of the tree must be
saved into memory to expand the next level
• BFS needs lots of time if the solution is far away from the root
node - 23What advantages/ disadvantages are there to DFS?• Advantage:
• DFS requires very less memory as it only needs to store a stack of
the nodes on the path from root node to the current node
• It takes less time to reach to the goal node than BFS algorithm (if it
traverses in the right path)
• Disadvantage:
• There is the possibility that many states keep re-occurring, and there
is no guarantee of finding the solution
• DFS algorithm goes for deep down searching and sometimes it may
go to the infinite loop - 24When to use DFS vs BFS?• Are you looking for the shortest path?
• If so, breadth-first search may be better as it will find the shortest path first.
• Is memory likely to be a problem?
• Depth-first search generally requires much less memory.
• Do you want to find a solution quickly?
• If so, the choice of algorithm gets complex!
• Depth-first search may be faster if there are many paths that lead to solution states, but
all are quite long.
• Breadth-first search may be faster if there is one short path to the target, but within a
large and deep search space - 25What kind of role does search algorithms have in information retrieval within search engines?Crawling:
Process where search engine bots (crawlers or spiders) discover web pages by following links and checking sitemaps.
Crawlers revisit pages periodically to detect updates.
Indexing:
Organizes and stores crawled data in a structured database for quick retrieval.
Analyzes content, metadata, and links to identify keywords, topics, and other relevant details.
Handles duplicate content and identifies languages and entities.
Query Processing
When a user submits a search query, the search engine processes it to understand what the user is looking for.
Ranking
After processing the query, the search engine evaluates how relevant each indexed webpage is to the user's search.
Displaying Results
The search engine displays ranked results, showing the most relevant pages at the top. This is the Search Engine Results Page (SERP)
User Feedback and Learning
Search engines monitor user interactions to improve future results.
If users often click on or spend more time on certain results, the search engine considers these results more relevant and adjusts rankings accordingly - 26Can u explain the travelling salesman problem?The travelling salesman problem (also called
the travelling salesperson problem or TSP) asks the
following question: "Given a list of cities and the distances
between each pair of cities, what is the shortest possible
route that visits each city exactly once and returns to the
origin city?
Informações
- Descrição
- FlashCards sobre Sem título, criado por Jessica Elias em 04-12-2024.
- Criado por
- Jessica Elias
- Visualizações
- 1
- Avaliação