Trending March 2024 # Airpods Pro 2 Have Everything You Need – But There’s One Feature Apple Hid # Suggested April 2024 # Top 10 Popular

You are reading the article Airpods Pro 2 Have Everything You Need – But There’s One Feature Apple Hid updated in March 2024 on the website Hatcungthantuong.com. We hope that the information we have shared is helpful to you. If you find the content interesting and meaningful, please share it with your friends and continue to follow and support us for the latest updates. Suggested April 2024 Airpods Pro 2 Have Everything You Need – But There’s One Feature Apple Hid

Apple announced AirPods Pro 2 during its “Far Out” event. While everyone keeps talking about the new Dynamic Island on iPhone 14 Pro series, there’s one product the Cupertino company outdid itself with – and it’s AirPods Pro 2. AirPods Pro 2 have everything you need – and more – but Apple has yet to announce one more feature this product could offer.

AirPods Pro 2 might look the same at first glance, but Apple rebuilt these wireless earbuds. While we yet have to see how it improved in real life, the company’s promises are impressive.

Personally, the original AirPods Pro fit just fine in my ears. Active Noise Cancellation is amazing, and these are my favorite wireless earbuds when I’m riding on the subway, taking the bus, or needing to disconnect from the world. Apple has claimed the new H2 chip can cancel noise up to two times better; this feature is something I’m looking forward to testing.

If ANC has improved, then Transparency Mode also has. Apple is now calling it Adaptive Transparency, which the company said can “reduce loud environmental noise – like a passing vehicle siren, construction tools, or even loud speakers at a concert.”

In addition, AirPods Pro 2 has another long-awaited feature: touch control. Users can finally lower or increase the volume by sliding their fingers on the stems of their AirPods.

While Apple had tweaked the first AirPods Pro case to work with the Find My app, the company has finally added the U1 chip in its MagSafe Charging Case, making it easier than ever to locate the wireless earbuds – and also make it play a sound, thanks to its built-in speakers.

Last but not least, the company is adding extra small ear tips so AirPods Pro 2 can fit more ears. With all that in mind, there’s one last thing that Apple didn’t mention during the keynote but could mean that this product could be even better than we expected.

AirPods Pro 2 offers Bluetooth 5.3 support, could bring new codec support, and “lossless” playback

A couple of months ago, Bluetooth SIG announced that it had finalized the LE Audio Spec, meaning companies could make use of this technology.

The Low Complexity Communication Codec – LC3 for short – can transmit at much lower bitrates without dropping the audio quality we currently see with Bluetooth’s standard – and the good part is that Apple is already testing it with AirPods Max.

A developer was able to activate this codec with the latest AirPods beta firmware, saying it’s improved the audio quality on calls. But something was missing on AirPods’ hardware: Bluetooth 5.2 support.

Thankfully, AirPods Pro 2 and iPhone 14 offer it, meaning Apple can not only support a higher bitrate with these products but also offer some kind of lossless streaming with Apple Music.

This new codec could let you listen to songs, watch movies, and more, with several wireless earbuds connected at the same time. Currently, you can only connect up to two AirPods from the same source.

While Apple seems to yet test this feature, the company could announce a revolution in audio streaming thanks to this new codec and this new hardware.

Wrap up

While I’m looking forward to Apple expanding AirPods Pro 2 capabilities, I’d also say you should not expect the company to officially announce support for this new codec. For example, the new Apple TV 4K could offer 120Hz refresh rate thanks to HDMI 2.1 support, but after a year, the company hasn’t updated the set-top box to offer this technology to users.

FTC: We use income earning auto affiliate links. More.

You're reading Airpods Pro 2 Have Everything You Need – But There’s One Feature Apple Hid

Everything You Need To Know About The Apple Watch Ultra

We may earn revenue from the products available on this page and participate in affiliate programs. Learn more ›

Bigger and tougher

The Apple Watch Ultra is bigger and more durable than the Apple Watch Series 8. Apple

Apple Watch Ultra is big. While the difference between the 45mm chassis of the Watch Series 8 and the 49mm Watch Ultra may not sound substantial, it should feel positively huge to standard Apple Watch users. Keep in mind: Apple expanded the case size by 1mm with the Watch Series 7, and that made a very noticeable difference. 

It’ll also have a much thicker chassis to incorporate new components, including a larger, louder speaker and a three-microphone array to improve voice clarity when making calls on the watch in less-than-ideal conditions. The Watch Ultra only comes in one hardware configuration, which includes cellular connectivity, so the expectation is that people will want to use the Watch Ultra to make calls at any time.

Presumably, the larger case also allowed Apple to give the Watch Ultra a bigger battery, which it estimates will last up to 36 hours on a single charge, or up to 60 hours with a low-power feature (available later in the fall).

The Apple Watch Ultra has a new “Action” button and a redesigned Digital Crown. Apple

The redesigned watch will also feature some design tweaks for the sake of durability, and usability in extreme conditions. The titanium case extends up to cover the edges of the sapphire crystal display to minimize cracked edges. The Watch Ultra is rated to operate on-wrist at temperatures as low as minus 4 degrees Fahrenheit, or as high as 131 F. It’s also IP6X and MIL-STD-810H certified—a military-grade durability rating used for many “rugged” tech products—indicating it’s prepared for some conditions, including rain, humidity, immersion in sand and dust, freezing, shock, and vibration, among others.

The buttons—yes, plural—are also getting an overhaul. The Digital Crown is larger and features grooved notches to make it easier to manipulate with a gloved hand. There’s also a second input: a large customizable “Action” button, that will allow you to start tracking workouts and perform other functions quickly. For example, triathletes can switch from running to cycling to swimming by simply pressing the button.

Last, but not least, Apple has created three new, activity-specific Apple Watch Ultra bands—the stitch-free hook-clasped Alpine Loop Band, the wetsuit-ready rubber Ocean Band, and the ultralight stretch Trail Loop band.

Built for survival

The new compass app allows you to set waypoints to help you find your way back to your camp or car. Apple

The Apple Watch Ultra offers some specialized features, many of which seem designed with safety and survival for hikers and climbers in mind. It uses a more precise “dual-frequency” GPS tracking that allows the watch to maintain tracking when you’re surrounded by tall structures or mountains.

As part of watchOS 9, the Watch Ultra will feature a redesigned version of the compass app that allows you to set waypoints, like your home, your camp, or your car, and allow you to orient yourself in relation to those locations. It will also be able to use a feature called backtrack that can use GPS to create a path retracing your steps in real-time. If you find yourself fully lost or hurt, the larger speaker can now play an ultra-loud 86-decibel siren that sends a distinctive SOS alarm (audible up to 600 feet away).

During the day, the display is brighter, up to 2000 Nits, which should make it easier to see regardless of glare. It also features a night mode, which turns the whole interface red, making it easier to see without interfering with your own night-adjusted vision.

Diver’s delight

The Apple Watch Ultra also seems to be an especially useful tool for divers. It’s waterproof up to 100 meters (WR100) and has an EN13319 depth gauge certification for diving accessories. Using a new depth app, you’ll be able to see your depth, time underwater, and max depth. In conjunction with an upcoming app, Oceanic+, the Watch Ultra will reportedly work as an effective dive computer, letting you plan and share dive routes and providing safety stop guidance.

Plus the best of Apple Watch Series 8 and watchOS 9

In addition to all of its exclusive changes, the Apple Watch Ultra will feature all of the upgrades in the upcoming Apple Watch Series 8. Most notably, that means new motion sensors that can detect if you get in a car crash and automatically call for help. They include a gyroscope and a highly sensitive accelerometer. Even the Watch Ultra’s built-in barometer plays a role in detecting crashes by detecting pressure changes typically associated with airbag deployment. There is also a temperature sensor that improves menstrual cycle tracking and enables ovulation tracking through the Health app (information Apple stressed is encrypted on the watch and only accessible with a user’s passcode/Touch ID/Face ID).

What does all this mean?

Apple will sell three activity-focused bands for the Apple Watch Ultra: The Trail Loop, the Alpine Loop, and the Ocean Band. Apple

At a glance, the people who should get most excited are iPhone-using fans of multisports smartwatches from brands like Garmin and Suunto. Those brands already make watches with many of these features, but their flagship watches cost even more than the $799 Apple Watch Ultra and don’t offer the same level of connectivity and convenience as an Apple Watch and iPhone working in sync.

The question remains: Is the Apple Watch Ultra worth buying? We will hopefully get our hands on the Apple Watch Ultra in the coming weeks, so we’ll have a full review with our thoughts on whether or not it’s worth that higher price. In the meantime, the Apple Watch Ultra is available on Amazon for $799.

How To Put Airpods Pro 2 Into Lost Mode

No matter how calculative you are, there could come a time when you may get caught off guard due to something or the other. That’s why life is more unpredictable than any other thing. So, I can understand what you are going through after having lost the brand-new AirPods Pro 2. The good thing is Apple offers a pretty effective way to put AirPods Pro 2 in the lost mode in order to help you get them back. Follow along to learn how it works: 

How to Enable Lost Mode for AirPods Pro 2 (Quick Guide)

Before going ahead with the guide, let’s briefly understand what is lost mode and how it works!

What is Lost Mode and How Does It Work? 

Lost Mode is designed to help you find your lost Apple devices including AirPods, iPhone, Apple Watch, and more. When you enable Lost Mode for your AirPods Pro 2, AirPods Pro 1, AirPods 3, or AirPods Max, Apple allows you to share a message with your phone number or email address. Hence, if anyone finds your lost AirPods, your custom message appears on their device. As a result, it becomes easier for the finder to contact you.

Play a Sound on Your AirPods Pro 2

Before turning on the lost mode, I would recommend you play a sound on your AirPods Pro 2. The second generation of AirPods Pro comes with a built-in speaker. Therefore, you can play a sound on the earphones if you ever find them missing.

If the earphones are near any of your Apple devices and are connected to Bluetooth, it should help you easily find them. Notably, you can also play a sound on the charging case of your second-generation AirPods Pro 2. So, give it a try first up. 

1. To get started, open the Find My app on your iPhone. 

2. Now, hit the Devices tab that shows at the bottom of the screen. 

3. Next, choose your AirPods from the list of devices. 

4. Next up, tap on Play Sound to play a sound on the AirPods. Note that the sound gets louder gradually.

Image credit: Apple

Turn on Lost Mode for Your AirPods Pro

1. On your iPhone or iPad, launch the Find My app. 

2. Now, tap on the Devices tab that appears at the bottom of the screen. 

3. Next, select your AirPods from the list of devices. 

4. Next up, scroll down to Mark as Lost section and hit Activate. 

5. After that, you need to follow the onscreen instructions if you want your contact information to display for your missing AirPods.

6. Finally, tap on Activate to confirm that you want to enable lost mode for your AirPods Pro. 

Image credit: Apple

Note: 

Note that if your AirPods are out of range or they need to charge, in this situation you will see their last known location, “No location found,” or “Offline.” 

In this scenario, you won’t be able to play a sound to find them. Though may be able to get directions to the location where your lost AirPods were last connected. 

It’s also important to note that if your AirPods ever come back online, you will get a notification on your iPhone, iPad, or Mac.

For an extra layer of security, you should enable the separation alert on your iPhone to get alerts if you leave your AirPods behind. It can play a vital role in preventing your AirPods from being lost.

If your AirPods are still missing and you are unable to find them, contact Apple Support at the earliest. Go to the AirPods setting (from the Settings app on your iPhone) and get the serial number. Then, contact Apple Support for a replacement.

Wrapping Up…

That’s pretty much it! So that’s how you can turn on the lost mode for your AirPods Pro 2. As already stated above, you can follow the same process to enable this mode for other AirPods variants including the AirPods Max.

While I do wish you may never need to try this feature, you never know what may happen tomorrow. Hence, it pays to be prepared rather than run for the cover at the eleventh hour. 

With iOS 16, Apple has introduced a dedicated setting for AirPods that appears right under your Apple ID banner. From a user experience perspective, it seems to be a welcome change. Thus, you will no longer need to dig into the Bluetooth setting to customize your AirPods.

Read more: How to Take a Screenshot With a Quick Tap in iOS 16 on iPhone

Apple Macbook Pro With M1 Pro, M1 Max Chip: All You Need To Know

About a year ago, Apple caused a stir in the computing world by introducing Apple Silicon M1 chips for its MacBook lineup. It announced a 2-year plan to transition away from Intel chips, and it’s following through with the latest M1 Pro and M1 Max chips.

These groundbreaking chips were revealed at the Unleashed event and boast up to a 10-core CPU, 32-core GPU, 64GB of unified memory, ProRes acceleration, and industry-leading power efficiency. What does all this mean in practicality, and what’s the difference between M1 Pro and M1 Max? Keep reading for all the details about Apple’s M1 lineup.

Source: Apple

What is the M1 Pro chip?

The successor of the M1 chip, the M1 Pro, is based on a 5nm architecture with 22.7 billion transistors, more than double the amount on M1. It features up to 10 CPU cores, including eight high-performance cores, for 70% faster performance than M1.

Further, M1 Pro has a GPU with up to 16 cores, eight more than M1, for 2x faster graphics. Coupled with a 16-core Neural Engine, a ProRes accelerator, and a powerful media engine, it lets you edit multiple instances of both 4K and 8K videos.

You can configure M1 Pro with up to 32GB of unified RAM, with up to 200GB/s of memory bandwidth, which is a 2x of M1 chip. M1 Pro supports up to two external displays.

What is the M1 Max chip? 

To complete the M1 family, the M1 Max is Apple’s largest chip yet, with 57 billion transistors and the same 10-core CPU and 16-core Neural Engine as the M1 Pro. The GPU can feature up to 32 cores. 

It also has an enhanced media engine and two ProRes accelerators, making it the ideal chip for anyone working with high-definition video.

Lastly, it offers up to 64GB of integrated RAM and doubles the memory interface of the M1 Pro with a bandwidth of up to 400 Gb/s. It also supports up to four external displays, so it is the ultimate choice for professionals working with games and graphics.

M1 Pro and M1 Max vs. M1 chip: What’s the difference?

The ARM-based M1 was Apple’s first custom chip for Macs. It proved faster than both Intel and AMD’s x86 processors while drawing far less power. Following that revolutionary move, Apple is taking a two-pronged approach with M1 Pro and M1 Max, the latest chips powering the new 14- and 16-inch MacBook Pros.

M1 tops out at 16GB of RAM, which is not enough for many pro users. The M1 Pro and Max support up to 64GB of RAM, making them the ideal choice for pro users. The table below shows an overall comparison of M1 vs. M1 Pro vs. M1 Max. 

What are the benefits of Apple’s M1 Pro and M1 Max chips?

CPU performance 

The M1 Pro and M1 Max’s new 10-core CPU with eight high-performance cores and two high-efficiency cores delivers stunning performance. Apple has shared that compared to the latest 8-core PC laptop chip, M1 Pro and M1 Max deliver up to 1.7x more CPU performance at the same power level.

Moreover, they match the PC chip’s peak performance using up to 70 percent less power. This means that even the most demanding tasks, like high-resolution graphics and video editing, are a breeze.

The graph below visually captures this blazing-fast performance. 

Source: Apple

GPU performance 

M1 Pro’s GPU can be configured up to 16-core and is up to 2x faster than M1. It’s also 7x faster than the integrated graphics on the latest 8-core PC laptop chip. 

Further, M1 Pro achieves greater performance while consuming up to 70 percent less power than a powerful discrete GPU for PC notebooks. At the same time, the M1 Pro can be configured with up to 32GB of fast unified memory, with up to 200GB/s of memory bandwidth, enabling creatives like 3D artists and game developers to do more on the go than ever before.

M1 Max takes this even further with a massive 32-core GPU for up to four times faster graphics performance than M1. The GPU delivers a similar performance to a high-end GPU in a compact pro PC laptop while consuming up to 40 percent less power. This means less heat is generated, the fans run quietly, and battery life is enhanced. 

Apple has shared the graph below to indicate M1 Max GPU performance. 

Source: Apple

Media Engine with ProRes

Moreover, it has dedicated acceleration for the ProRes professional video codec. This allows playback of multiple streams of 4K and 8K ProRes video while using very little power. 

M1 Max takes it further with up to 2x faster video encoding than M1 Pro and two ProRes accelerators. Powered by this technology, the new MacBook Pro can transcode ProRes video in Compressor up to 10x faster than the previous-generation 16-inch MacBook Pro. That’s a fantastic step-up for video creators. 

Other features 

Apart from the impressive performance enhancements, the M1 Pro and M1 Max boast several other features that can take pro workflows to the next level:

The custom image signal processor and the Neural Engine use computational video to enhance image quality for sharper video and more natural-looking photos in the built-in camera. 

Ability to support multiple external displays 

Additional integrated Thunderbolt 4 controllers for even more I/O bandwidth.

Best-in-class security

The 2023 MacBook Pro with M1 Pro and M1 Max chips

The latest M1 Pro and M1 Max chips currently come exclusively on Apple’s all-new 2023 MacBook Pro. This beast of a device offers 14- and 16-inch display options with a new design that brings back many of the most-loved external features of earlier MacBook Pros with the mega internal power of the M1 Pro and M1 Max system-on-a-chip architectures. 

Apple has parted ways with the Touch Bar in favor of the more user-friendly physical function buttons. It’s also brought back the SDXC card slot and an HDMI port that will keep graphics and video professionals happy.

On the battery life front, Apple says the 14-inch MacBook Pro (2024) is rated to deliver up to 17 hours of video playback on a single charge, while its 16-inch counterpart delivers up to 21 hours of video playback.

You can charge the new MacBook Pro with a MagSafe charging cable with MagSafe 3, unlike the existing MacBook Pro models that offer charging via USB-C. However, you can still use Thunderbolt 4 for charging using a USB Type-C cable.

Moreover, the display features a notch housing a 1080p webcam. This design change helps reduce bezels and increase screen real estate.

As for audio, there’s a powerhouse six-speaker sound system with two tweeters and four force-canceling woofers for audio. There is also Dolby Atmos and spatial audio support for a superior surround sound experience. 

Overall, it’s one of the most significant updates to the Mac lineup in recent years. It’s the best device for power users in creative fields like filmmaking or game development. 

M1 Pro and M1 Max: Frequently asked questions

You might also like to read:

Author Profile

Mehak

Mehak has a master’s degree in communication and over ten years of writing experience. Her passion for technology and Apple products led her to iGeeksBlog, where she specializes in writing product roundups and app recommendations for fellow Apple users. When not typing away on her MacBook Pro, she loves being lost in a book or out exploring the world.

Everything You Need To Know About Scikit

Introduction

Scikit-learn is one Python library we all inevitably turn to when we’re building machine learning models. I’ve built countless models using this wonderful library and I’m sure all of you must have as well.

There’s no question – scikit-learn provides handy tools with easy-to-read syntax. Among the pantheon of popular Python libraries, scikit-learn ranks in the top echelon along with Pandas and NumPy. These three Python libraries provide a complete solution to various steps of the machine learning pipeline.

I love the clean, uniform code and functions that scikit-learn provides. It makes it really easy to use other techniques once we have mastered one. The excellent documentation is the icing on the cake as it makes a lot of beginners self-sufficient with building machine learning models.

The developers behind scikit-learn have come up with a new version (v0.22) that packs in some major updates. I’ll unpack these features for you in this article and showcase what’s under the hood through Python code.

Note: Looking to learn Python from scratch? This free course is the perfect starting point!

Table of Contents

Getting to Know Scikit-Learn

A Brief History of Scikit-Learn

Scikit-Learn v0.22 Updates (with Python implementation)

Stacking Classifier and Regressor

Permutation-Based Feature Importance

Multi-class Support for ROC-AUC

kNN-Based Imputation

Tree Pruning

Getting to Know Scikit-Learn

This library is built upon the SciPy (Scientific Python) library that you need to install before you can use scikit-learn. It is licensed under a permissive simplified BSD license and is distributed under many Linux distributions, encouraging academic and commercial use.

Overall, scikit-learn uses the following libraries behind the scenes:

NumPy: n-dimensional array package

SciPy: Scientific computing Library

Matplotlib:  Plotting Library

iPython: Interactive python (for Jupyter Notebook support)

SymPy: Symbolic mathematics

Pandas: Data structures, analysis, and manipulation

Lately, scikit-learn has reorganized and restructured its functions & packages into six main modules:

Classification: Identifying which category an object belongs to

Regression: Predicting a continuous-valued attribute associated with an object

Clustering: For grouping unlabeled data

Dimensionality Reduction: Reducing the number of random variables to consider

Model Selection: Comparing, validating and choosing parameters and models

Preprocessing: Feature extraction and normalization

scikit-learn provides the functionality to perform all the steps from preprocessing, model building, selecting the right model, hyperparameter tuning, to frameworks for interpreting machine learning models.

Scikit-learn Modules (Source: Scikit-learn Homepage)

A Brief History of Scikit-learn

Scikit-learn has come a long way from when it started back in 2007 as scikits.learn. Here’s a cool trivia for you – scikit-learn was a Google Summer of Code project by David Cournapeau!

This was taken over and rewritten by Fabian Pedregosa, Gael Varoquaux, Alexandre Gramfort and Vincent Michel, all from the French Institute for Research in Computer Science and Automation and its first public release took place in 2010.

Since then, it has added a lot of features and survived the test of time as the most popular open-source machine learning library across languages and frameworks. The below infographic, prepared by our team, illustrates a brief timeline of all the scikit-learn features along with their version number:

The above infographics show the release of features since its inception as a public library for implementing Machine Learning Algorithms from 2010 to 2023

Today, Scikit-learn is being used by organizations across the globe, including the likes of Spotify, JP Morgan, chúng tôi Evernote, and many more. You can find the complete list here with testimonials I believe this is just the tip of the iceberg when it comes to this library’s popularity as there will a lot of small and big companies using scikit-learn at some stage of prototyping models.

The latest version of scikit-learn, v0.22, has more than 20 active contributors today. v0.22 has added some excellent features to its arsenal that provide resolutions for some major existing pain points along with some fresh features which were available in other libraries but often caused package conflicts.

We will cover them in detail here and also dive into how to implement them in Python.

Scikit-Learn v0.22 Updates

Along with bug fixes and performance improvements, here are some new features that are included in scikit-learn’s latest version.

Stacking Classifier & Regressor

Stacking is an ensemble learning technique that uses predictions from multiple models (for example, decision tree, KNN or SVM) to build a new model.

This model is used for making predictions on the test set. Below is a step-wise explanation I’ve taken from this excellent article on ensemble learning for a simple stacked ensemble:

The base model (in this case, decision tree) is then fitted on the whole train dataset

This model is used to make final predictions on the test prediction set

The mlxtend library provides an API to implement Stacking in Python. Now, sklearn, with its familiar API can do the same and it’s pretty intuitive as you will see in the demo below. You can either import StackingRegressor & StackingClassifier depending on your use case:

from

sklearn.linear_model

import

LogisticRegression

from sklearn.ensemble import RandomForestClassifier from chúng tôi import DecisionTreeClassifier

from

sklearn.ensemble

import

StackingClassifier

from

sklearn.model_selection

import

train_test_split

X

,

y

=

load_iris

(

return_X_y

=

True

)

estimators

=

[

(

'rf'

,

RandomForestClassifier

(

n_estimators

=

10

,

random_state

=

42

)),

(

'dt'

,

DecisionTreeClassifier

(

random_state

=

42

)

)

]

clf

=

StackingClassifier

(

estimators

=

estimators

,

final_estimator

=

LogisticRegression

()

)

X_train

,

X_test

,

y_train

,

y_test

=

train_test_split

(

X

,

y

,

stratify

=

y

,

random_state

=

42

)

clf

.

fit

(

X_train

,

y_train

)

.

score

(

X_test

,

y_test

)

Permutation-Based Feature Importance

As the name suggests, this technique provides a way to assign importance to each feature by permuting each feature and capturing the drop in performance.

But what does permuting mean here? Let us understand this using an example.

Let’s say we are trying to predict house prices and have only 2 features to work with:

LotArea – (Sq Feet area of the house)

YrSold (Year when it was sold)

The test data has just 10 rows as shown below:

Next, we fit a simple decision tree model and get an R-Squared value of 0.78. We pick a feature, say LotArea, and shuffle it keeping all the other columns as they were:

Next, we calculate the R-Squared once more and it comes out to be 0.74. We take the difference or ratio between the 2 (0.78/0.74 or 0.78-0.74), repeat the above steps, and take the average to represent the importance of the LotArea feature.

We can perform similar steps for all the other features to get the relative importance of each feature. Since we are using the test set here to evaluate the importance values, only the features that help the model generalize better will fare better.

Earlier, we had to implement this from scratch or import packages such as ELI5. Now, Sklearn has an inbuilt facility to do permutation-based feature importance. Let’s get into the code to see how we can visualize this:



As you can see in the above box plot, there are 3 features that are relatively more important than the other 4. You can try this with any model, which makes it a model agnostic interpretability technique. You can read more about this machine learning interpretability concept here.

Multiclass Support for ROC-AUC

The ROC-AUC score for binary classification is super useful especially when it comes to imbalanced datasets. However, there was no support for Multi-Class classification till now and we had to manually code to do this. In order to use the ROC-AUC score for multi-class/multi-label classification, we would need to binarize the target first.

Currently, sklearn has support for two strategies in order to achieve this:

from sklearn.datasets import load_iris  from sklearn.ensemble import RandomForestClassifier from sklearn.metrics import roc_auc_score X, y = load_iris(return_X_y=True) rf = RandomForestClassifier(random_state=44, max_depth=2) rf.fit(X,y) print(roc_auc_score(y, rf.predict_proba(X), multi_class='ovo'))

Also, there is a new plotting API that makes it super easy to plot and compare ROC-AUC curves from different machine learning models. Let’s see a quick demo:

from

sklearn.model_selection

import

train_test_split

from

sklearn.svm

import

SVC

from

sklearn.metrics

import

plot_roc_curve

from

sklearn.ensemble

import

RandomForestClassifier

from

sklearn.datasets

import

make_classification

import

matplotlib.pyplot

as

plt

X

,

y

=

make_classification

(

random_state

=5

)

X_train

,

X_test

,

y_train

,

y_test

=

train_test_split

(

X

,

y

,

random_state

=

42

)

svc

=

SVC

(

random_state

=

42

)

svc

.

fit

(

X_train

,

y_train

)

rfc

=

RandomForestClassifier

(

random_state

=

42

)

rfc

.

fit

(

X_train

,

y_train

)

svc_disp

=

plot_roc_curve

(

svc

,

X_test

,

y_test

)

rfc_disp

=

plot_roc_curve

(

rfc

,

X_test

,

y_test

,

ax

=

svc_disp

.

ax_

)

rfc_disp

.

figure_

.

suptitle

(

"ROC curve comparison"

)

plt

.

show

()

In the above figure, we have a comparison of two different machine learning models, namely Support Vector Classifier & Random Forest. Similarly, you can plot the AUC-ROC curve for more machine learning models and compare their performance.

kNN-Based Imputation

In kNN based imputation method, the missing values of an attribute are imputed using the attributes that are most similar to the attribute whose values are missing. The assumption behind using kNN for missing values is that a point value can be approximated by the values of the points that are closest to it, based on other variables.

The k-nearest neighbor can predict both qualitative & quantitative attributes

Creation of predictive machine learning model for each attribute with missing data is not required

Correlation structure of the data is taken into consideration

Scikit-learn supports kNN-based imputation using the Euclidean distance method. Let’s see a quick demo:

import

numpy

as

np

from

sklearn.impute

import

KNNImputer

X

=

[[4

,

6

,

np

.

nan

],

[

3

,

4

,

3

],

[

np

.

nan

,

6

,

5

],

[

8

,

8

,

9

]]

imputer

=

KNNImputer

(

n_neighbors

=

2

)

print

(

imputer

.

fit_transform

(

X

))

You can read about how kNN works in comprehensive detail here.

Tree Pruning

In basic terms, pruning is a technique we use to reduce the size of decision trees thereby avoiding overfitting. This also extends to other tree-based algorithms such as Random Forests and Gradient Boosting. These tree-based machine learning methods provide parameters such as min_samples_leaf and max_depth to prevent a tree from overfitting.

Pruning provides another option to control the size of a tree. XGBoost & LightGBM have pruning integrated into their implementation. However, a feature to manually prune trees has been long overdue in Scikit-learn (R already provides a similar facility as a part of the rpart package).

In its latest version, Scikit-learn provides this pruning functionality making it possible to control overfitting in most tree-based estimators once the trees are built. For details on how and why pruning is done, you can go through this excellent tutorial on tree-based methods by Sunil. Let’s look at a quick demo now:

from sklearn.ensemble import RandomForestClassifier from sklearn.datasets import make_classification X

,

y

=

make_classification

(

random_state

=

0

)

rf

=

RandomForestClassifier

(

random_state

=

0

,

ccp_alpha

=

0

)

.

fit

(

X

,

y

)

print

(

"Average number of nodes without pruning

{:.1f}

"

.

format

(

np

.

mean

([

e

.

tree_

.

node_count

for

e

in

rf

.

estimators_

])))

rf

=

RandomForestClassifier

(

random_state

=

0

,

ccp_alpha

=

0.1

)

.

fit

(

X

,

y

)

print

(

"Average number of nodes with pruning

{:.1f}

"

.

format

(

np

.

mean

([

e

.

tree_

.

node_count

for

e

in

rf

.

estimators_

])))

End Notes

The scikit-learn package is the ultimate go-to library for building machine learning models. It is the first machine learning-focused library all newcomers lean on to guide them through their initial learning process. And even as a veteran, I often find myself using it to quickly test out a hypothesis or solution I have in mind.

The latest release definitely has some significant upgrades as we just saw. It’s definitely worth exploring on your own and experimenting using the base I have provided in this article.

Related

Xiaomi 12S: Everything You Need To Know

Just as importantly, the trio of 12S phones unveiled in China – the 12S, 12S Pro, and 12S Ultra – are the first to bear a Leica logo, marking the two companies’ new long-term camera collaboration.

Here’s everything we know so far about the three phones, including their prospects of launching in the West.

When will the Xiaomi 12S launch worldwide?

Xiaomi revealed the three 12S phones in China on 4 July, but what about the rest of the world?

The Xiaomi 12 and 12 Pro took a couple of months to get a global launch, so on that timetable we’d expect to see the 12S arrive globally some time in September.

That’s assuming they’ll launch at all. Last year’s Mi 11 Ultra never officially reached the West at all, and Xiaomi has already confirmed to Tech Advisor that “The series will not be coming to the UK.”

There’s still a chance it hits some other markets outside China of course, though right now we’d consider it a long shot. Hopefully we’ll find out more soon.

How much do the 12S phones cost?

With a launch only in China, we only have Chinese prices for the phones so far, but here’s how they convert worldwide:

Xiaomi 12S – From ¥3,999 (around $600/£490/€570)

Xiaomi 12S Pro – From ¥4,699 (around $700/£580/€670)

Xiaomi 12S Ultra – From ¥5,999 (around $900/£740/€860)

What are the Xiaomi 12S specs?

Now that Xiaomi has released the 12S series in China, we know the specs for all three phones. Here’s how they stack up:

Xiaomi 12S specs

The regular 12S will be the smallest and cheapest of the phones in the series, but don’t hold that against it. It’s available in black, purple, green, and a new white finish.

Xiaomi

It uses the same size 6.28in display as the Xiaomi 12, keeping things nice and compact. As you’d expect, it will be a 120Hz AMOLED at a Full HD+ resolution – exactly the same as its predecessor.

There are still some major performance upgrades though. The most obvious it that, like all the 12S phones, it will use the Snapdragon 8+ Gen 1 chipset. That should result in some modest performance gains, and perhaps better battery life – though the actual cell is the same 4500mAh capacity, with 67W wired and 50W wireless charging.

The camera has also been upgraded. The main rear lens is still 50Mp, but it now uses the improved Sony IMX707 sensor – the same used in the 12 Pro earlier this year – with OIS. Like all three of the new phones, it also features Leica branding on the rear to reflect the camera company’s work helping Xiaomi to optimise image quality.

As with all phones in the series, the Xiaomi 12S launches with Android 12 and MIUI 13 installed.

Here are the full Xiaomi 12S specs:

6.28in, 120Hz, FHD+ AMOLED display

Qualcomm Snapdragon 8+ Gen 1

8/12GB RAM

128/256GB storage

50Mp, f/1.9 OIS main camera

13Mp, f/2.4 ultrawide camera

5Mp telemacro camera

32Mp selfie camera

4500mAh battery

67W wired charging

50W wireless charging

Gorilla Glass Victus

152.7 x 69.9 x 8.2 mm

180g

Xiaomi 12S Pro specs

As with the regular Xiaomi 12 line, the 12S Pro sits only a little above the 12S in the hierarchy, and is differentiated by the fact that it’s larger, and has better specced cameras and display.

Xiaomi

Like the 12S you’ll get the Snapdragon 8+ Gen 1, but with slightly improved performance thanks to a larger cooling area. The chipset choice is essentially the only key upgrade from the 12 Pro.

The bigger 6.73in E5 AMOLED display will likely be a big draw to many, backed up by a dynamic 1-120Hz refresh rate and high 2K resolution – though this is essentially unchanged from the 12 Pro.

You also get the same fast charging as the 12 Pro (120W wired, 50W wireless) and the same rear camera setup.

Here are the full specs:

6.73in, 1-120Hz LTPO, 2K AMOLED display

Qualcomm Snapdragon 8+ Gen 1

8/12GB RAM

256GB storage

50Mp, f/1.9 OIS main camera

50Mp, f/2.2 ultrawide camera

50Mp, f/1.9 2x zoom camera

32Mp selfie camera

4600mAh battery

120 wired charging

50W wireless charging

Gorilla Glass Victus

163.6 x 74.6 x 8.2 mm

205g

Xiaomi 12S Ultra specs

The 12S Ultra is a more different phone. That’s obvious immediately from the design – not only is the 12S Ultra larger, it also has a totally different circular rear camera module, surrounded by a 23K gold ring.

Xiaomi

Unlike last year’s Mi 11 Ultra there’s no space carved out for a mini rear display, and instead this bank of space is all about the actual camera – though we’ll get to that in a minute.

The phone comes in a vegan leather finish, available in two colours: Classic Black and Verdant Green. It’s also IP68-rated for water-resistance.

As for the display, you get a 6.73in 120Hz LTPO 2.0 AMOLED in a 2K resolution. As with the other two phones, the 12S Ultra uses the Snapdragon 8+ Gen 1 chipset.

The Ultra is also set to impress on battery. It features a 4860mAh battery with 67W wired charging and 50W wireless – so it actually has slower wired charging than the 12 Pro, but with a slightly larger battery. It also includes a proprietary Surge G1 battery management chipset.

Xiaomi

Now let’s get to the camera. While early reports had suggested the Ultra might be the debut of one of Samsung’s 200Mp camera sensors, in fact the phone uses a brand-new Sony sensor for its main camera, at aperture f/1.9. The new IMX989 is notable for being a 1in sensor – similar to the one equipped by Sony’s own Xperia Pro-I.

It’s paired with 48Mp ultrawide and 5x periscope lenses in essentially the same camera setup as the Mi 11 Ultra, outside of the upgraded main lens. The fourth rear lens is simply a ToF sensor for depth detection, while the selfie camera is be a 20Mp shooter.

6.73in, 120Hz LTPO 2.0, 2K AMOLED display

Qualcomm Snapdragon 8+ Gen 1

8/12GB RAM

256/512GB storage

50Mp, f/1.9 OIS main camera

48Mp, f/2.2 ultrawide camera

48Mp, f/4.1 OIS 5x periscope camera

32Mp selfie camera

4860mAh battery

67W wired charging

50W wireless charging

IP68 rating

Gorilla Glass Victus

Update the detailed information about Airpods Pro 2 Have Everything You Need – But There’s One Feature Apple Hid on the Hatcungthantuong.com website. We hope the article's content will meet your needs, and we will regularly update the information to provide you with the fastest and most accurate information. Have a great day!