588

Sollicitatievragen voor Applied Scientist gedeeld door sollicitanten

Meest gestelde sollicitatievragen

Sorteren: Relevantie|Populair|Datum
Er werd een Applied Scientist gevraagd...2 september 2016

How to write a function to make a biased coin from a fair coin and vice versa

4 antwoorden

a given coin is a function that randomly outputs 0 or 1. a fair coin outputs 0 with a probability of 0.5. altering a coin can be done by wrapping its function into another function that alters its result by making use of some random to override the original output with a new given probability. define a function that creates a biased coin (which is a function), given the original coin, the bias amount and the bias direction (towards 0 or 1). This allows us to solve the second part of the question: creating a fair coin from a biased coin: first we determine the bias degree of the biased coin by throwing it for say 100 times and then we create an altered coin to overcome that bias, resulting in a fair coin. Minder

Let we need head : tail probability as a:b where a and b are relatively prime integers. then we toss coin n times such that 2^n &gt;= (a+b). we convert the output into a number using standard binary conversion with the head as 1 and tail as 0(HTT, for example, will become 4). If the number is less than a, we return head, else if the number is less than (a+b) we return tail else we repeat the process. Example:- Lets us create coin with 2:1 probability of Head:tail We toss coin 2 times If its a TT or TH we return head, if its HT we return tail, if its HH we again repeat the toss. generating equal head and tail from biased coin:- toss coin two times. If its a TH return tail if its HT return head else toss again. Minder

You use uniform distribution to generate a uniformly distributed random number, and you use the odds(weights) as decision boundary to decide the results. For example if a biased coin weights is given by 0.2 and 0.8, then you generate a random number between 0 and 1, if the random number between [0, 0,2) you return . head, if the number is between [0.2, 1) you return tail. Same method also applies to multi-classes weighted sampling Minder

Meer reacties weergeven

• What are the typical Greek symbols used in Q-Learning? • What does Alpha typically represent? • What does Gamma typically represent? • What does Epsilon typically represent? • What is Greedy-Epsilon? • How does a High Alpha versus a Low Alpha impact the model? • What is the Exploration-Exploitation Tradeoff? • What is a Decay Structure? • What is important about a Decay Structure? • How could we apply reinforcement learning to Alexa/Echo which would add functionality? • How would you implement this? • What kind of reward structure would you use? • Why would you use that reward structure? • Tell me about a time when you were not able to complete all parts of a task? • Tell me about a time you not only met expectations but exceeded them?

4 antwoorden

Q-Learning delta &amp; gamma

alpha: learning rate

Gamma: discount factor)

Meer reacties weergeven

How to test a fair coin?

3 antwoorden

You should perform a hypothesis testing, in this case the null hypothesis is p(head) = p(tail) which is 0.5 each. Then you run experiment to collect results, calculate the p value if null hypothesis is true. If p value is smaller than 5 (in most cases), you can't reject hypothesis. If p value is bigger than 5, then the null hypothesis is False, then it's a unfair coin Minder

Correct the answer i gave above, if p value is smaller than 5 then you can reject the null hypothesis and vice versa Minder

throw the coin for say 100 times. it should land 50/50 overall

pros and cons of random forest and why

3 antwoorden

Pros Easy to interpret Handles both categorical and continuous data well. Works well on a large dataset. Not sensitive to outliers. Non-parametric in nature. Cons These are prone to overfitting. It can be quite large, thus making pruning necessary. It can’t guarantee optimal trees. It gives low prediction accuracy for a dataset as compared to other machine learning algorithms. Calculations can become complex when there are many class variables. High Variance(Model is going to change quickly with a change in training data) Minder

Random Forest Pros: Robust to outliers. Works well with non-linear data. Lower risk of overfitting. Runs efficiently on a large dataset. Better accuracy than other classification algorithms. Cons: Random forests are found to be biased while dealing with categorical variables. Slow Training. Not suitable for linear methods with a lot of sparse features Minder

the first one is pros and cons of decision tree

The interviewer was team manager and explained what his team is doing. After that, he asked me on my previous projects. He explained one of the projects his team is working on and asked me how to solve some specific problems related to the project.

3 antwoorden

Got an offer.

When did you hear from them? Did you apply for internship position or full time?

Have you got any response yet?

Using satellite imagery, how can you tell the difference between a corn field and a soy field?

2 antwoorden

1 - The color is probably different (this is where she claimed the 30 meters per pixel resolution, and they look the same in the pictures! so I had to add other ways). 2 - The harvest season is different and real time imagery can tell which one is which (she accepted this answer). Minder

The NASA and ESA imagery she referred to was most likely Landsat, which has a resolution of 30m, and Sentinel, which is 20m. These are publicly available datasets that both revisit the same points on the earth at regular intervals. The comparison to Apple Maps isn't appropriate, nor is Google maps. Their satellite layers are composite based maps based on multiple data sources, some of which are airborne sensors. But more importantly these are commercial data sources without open licenses like Landsat and Sentinel. Furthermore, land sensing instruments have many more bands that just the 3 bands that are available in a simple base composite base layer. Minder

Q: Given a function with inputs --an array with N randomly sorted numbers, and an int K, return output in an array with the K largest numbers. Q: 1. How does GMM/HMM work 2. Name some dimensional reduction method; I said PCA and we talked a bit about how PCA works and what's the physical intuiation 3. How K-means work, what kind of distance metric would you choose, what if different features have different dynamic range 4. How GMM works (EM algorithm)

2 antwoorden

A: The correct answer should be use heap or selection algorithm. But I tried using sorting, which would be more complex. Minder

Should I expect some ML system designs as well? Thank you for sharing your experience. Wondering have you ever tried Rooftop Slushie for your interview prep? I saw several posts similar to your experience on this community, which was really helpful for me so far. rooftopslushie.com Minder

1- Machine Learning concepts 2- Data Structure Case Study (Very easy) 3- Coding (Easy but they need hands-on exp) And Finally, Leadership (14 principles)

2 antwoorden

There is no right answer except for conceptual one. They need you to be natural and be good in what you do. I would suggest to do following: 1- Leadership 2- Coding 3- ML concepts Minder

Forgot to add, my feedback was mixed saying I need more Machine Learning knowledge (which I disagree) and hands-on coding (which killed my interview)! Minder

Do you have any experience about machine learning other than your research field? What is the probability of blah-blah...? Write a code to switch a k-th element with (N-k)-th element in a linked list of length N.

2 antwoorden

run a pointer k steps, store a reference of the node, start a second pointer, run both until the first pointer reaches the end of list, flip the stored kth element and first pointer. Minder

run a pointer k steps, store a reference of the node, start a second pointer, run both until the first pointer reaches the end of list, flip the stored kth element and first pointer. Minder

The interview questions were very specific to my previous experience in Sports Science and Research. A further question was how well I could integrate with software engineers as a Sports Scientist.

1 antwoorden

I gave a lot of detail regarding my experience and how said experience related to and/or would be beneficial to my as an employee at Kitman Labs. Minder

Weergave: 1 - 10 van 588 sollicitatievragen