On May 7, 2021, Colonial pipelines went through one of the biggest cyber attacks. The hackers broke into the computer system that eventually impacted the computerized equipment. The attack was a ransom ware attack, where the hackers named Darkside demanded a ransom of 75 bit coins ($4.4 million). Colonial pipelines accepted to pay the ransom money because that was the best option for the country.

Introduction:

      The largest pipeline system in the US for refined oil products is the colonial pipeline. The construction of the company began in 1962 and its headquarter is in Alpharetta, Georgia. On the coast of the Gulf of Mexico, an 8900km long pipeline terminates at the port of New York and New Jersey. Colonial pipeline delivers 10000000 gallons of gasoline, heating oil’s, refined petroleum oils, to businesses throughout the east and south coast. About 45% of the fuel consumed on the east coast is derived from colonial pipelines.

      Anyways colonial pipeline was under threat when a cyber-attack broke into the data of the company in 2021. On 29th April 2021, through a virtual private network (VPN) that the employees used to work remotely, the hackers got a hold of the password and the user name that helped them to access the information and data. It is only possible if the employee has used the same password to access another account that was hacked in the past.

Facts about the attack:

• The cyber attack was made open to the public on the 8th of May, 2021, at 12:30 pm.
• Colonial pipeline started to shut down their online system.
• While the data and information were restoring, the gas supply was limited.
• The Colonial pipeline involved the law and also took help from a third party during the recovery.
• People panicked because of this attack which created a shortage of gas in the country.
• The malicious attackers were named Darkside.
• Colonial pipelines paid a ransom of 75 bit coins ($4.4 million)
• Darkside lost access to information about the infrastructure and funds.
• The data was restored on the 13th of May, 2021, and the system started running online again.

YOLOv4: Optimal Speed and Accuracy of Object Detection

                                                       -MEETALI SONI

1. Introduction

The majority of CNN-based object detectors are largely applicable only for recommendation systems. For example, searching for free parking spaces via urban video cameras is executed by slow accurate models, whereas car collision warning is related to fast inaccurate models. Improving the real-time object detection accuracy enables using them not only for hint generating recommendation systems but also for stand-alone process management and human input reduction. Real-time object detector operation on conventional Graphics Processing Units (GPU) allows their mass usage at an affordable price. The most accurate modern neural networks do not operate in real-time and require a large number of GPUs for training with a large mini-batchsize. We address such problems by creating a CNN that operates in real-time on a conventional GPU, and for which training requires only one conventional GPU.

Figure 1: Comparison of the proposed YOLOv4 and other state-of-the-art object detectors. YOLOv4 runs twice faster than efficient with comparable performance. Improves YOLOv3’s AP and FPS by 10% and 12%, respectively.

The main goal of this work is designing a fast-operating speed of an object detector in production systems and optimization for parallel computations, rather than the low computation volume theoretical indicator.

• Related work
• Object Detection Models

A modern detector is usually composed of two parts, a backbone which is pre-trained on ImageNet, and a head which is used to predict classes and bounding boxes of objects. For those detectors running on GPU platform, their backbone could be VGG [68], ResNet [26], ResNeXt [86], or DenseNet [30]. For those detectors running on the CPU platform, their backbone could be Squeeze Net [31], Mobile Net [28, 66, 27, 74], or Shuffle Net [97, 53].

Partsof Object Detector

1. Input:Image, Patches, Image Pyramid
   1. Backbones:VGG16 [68], ResNet-50 [26], SpineNet [12], EfficientNet-B0/B7 [75], CSPResNeXt50 [81], CSPDarknet53 [81]
   1. Neck:
2. Additional blocks:SPP [25], ASPP [5], RFB [47], SAM [85]
3. Path-aggregation blocks:FPN [44], PAN [49], NAS-FPN [17], Fully-connected FPN, BiFPN [77], ASFF [48], SFAM [98]
   1. Heads:
4. Dense Prediction (one-stage): RPN [64], SSD [50], YOLO [61], RetinaNet [45] (anchor based) ◦ CornerNet [37], CenterNet [13], MatrixNet [60], FCOS [78] (anchor free)
5. Sparse Prediction (two-stage):Faster R-CNN [64], R-FCN [9], Mask RCNN [23] (anchor based) ◦ RepPoints [87] (anchor free)

Comparison of the proposed YOLOv4 and other state-of-the-art object detectors. The dashed line means only latency of model inference, while the solid line includes model inference and post-processing.

• Bag of Freebies

Usually, a conventional object detector is trained offline. Therefore, researchers always like to take this advantage and develop better training methods that can make the object detector receive better accuracy without increasing the inference cost. We call these methods that only change the training strategy or only increase the training cost a “bag of freebies.”The purpose of data augmentation is to increase the variability of the input images so that the designed object detection model has higher robustness to the images obtained from different environments.

In dealing with the problem of semantic distribution bias, a very important issue is that there is a problem of data imbalance between different classes, and this problem is often solved by hard negative example mining [72] or online hard example mining [67] in two-stage object detector.

But the example mining method does not apply to a one-stage object detector, because this kind of detector belongs to the dense prediction architecture. Therefore Lin et al. [45] proposed focal loss to deal with the problem of data imbalance existing between various classes. Another very important issue is that it is difficult to express the relationship of the degree of association between different categories with the one-hot hard representation.

To make this issue processed better, some researchers recently proposed IoU loss [90], which puts the coverage of predicted BBox area and ground truth BBox area into consideration. The IoU loss computing process will trigger the calculation of the four coordinate points of the BBox by executing IoU with the ground truth and then connecting the generated results into a whole code. Because IoU is a scale-invariant representation. Recently, some researchers have continued to improve IoU loss.

• Bag of Specials

For those plugin modules and post-processing methods that only increase the inference cost by a small amount but can significantly improve the accuracy of object detection, we call them “bag of specials”. Generally speaking, these plugin modules are for enhancing certain attributes in a model, such as enlarging receptive field, introducing attention mechanism, or strengthening feature integration capability, etc., and post-processing is a method for screening model prediction results.

Since multi-scale prediction methods such as FPN have become popular, many lightweight modules that integrate different feature pyramids have been proposed. The modules of this sort include SFAM [98], ASFF [48], and BiFPN [77]. The main idea of SFAM is to use the SE module to execute channel-wise level re-weighting on multi-scale concatenated feature maps.As for ASFF, it uses SoftMax as point-wise level reweighting and then adds feature maps of different scales. In BiFPN, the multi-input weighted residual connections are proposed to execute scale-wise level re-weighting, and then add feature maps of different scales.

Example of the Representation Chosen when Predicting Bounding Box Position and Shape Taken from YOLO9000: Better, Faster, Stronger

The post-processing method commonly used in deep learning-based object detection is NMS, which can be used. The DIoU NMS [99] developers’ way of thinking is to add the information of the center point distance to the BBox screening process based on soft NMS. It is worth mentioning that, since none of the above postprocessing methods directly refer to the captured image features, post-processing is no longer required in the subsequent development of an anchor-free method.

• Advantages
• It’s incredibly fast and can process 45 frames per second to 150 frames per second.
• YOLO also understands generalized object representation.
• The network can generalize the image better.

• Disadvantages
• Comparatively low recall and more localization error compared to Faster R_CNN.
• Struggles to detect close objects because each grid can propose only 2 bounding boxes.
• Struggles to detect small objects.
• Conclusion

We offer a state-of-the-art detector that is faster (FPS) and more accurate (MS COCO AP50…95 and AP50) than all available alternative detectors. The detector described can be trained and used on a conventional GPU with 8-16 GB-VRAM this makes its broad use possible. The original concept of one-stage anchor-based detectors has proven its viability.

“Health is not valued till sickness comes”

Health is “God’s gift to us”. Health refers to the physical and mental state of a human being. To stay healthy is not an option but a necessity to live a happy life. The basic laws of good health are related to the food we eat, the amount of physical exercise we do, our cleanliness, rest, and relaxation. A healthy person is normally more confident, self-assured, sociable, and energetic. A healthy person views things clearly, calmly, and without prejudice. 

Introduction :

We live in a super-fast age. The Internet has shrunk the world dramatically and people are connected 24×7. Multitasking is the order of the day, as we struggle to fulfill our responsibilities for everyone in life. In this fight, we often forget to spare time for ourselves. The stress levels continue to build up until one day a major collapse may make us realize that in all this hectic activity, we have forgotten to take care of one important thing – our health. As a we spend days shuttling between hospital and home, put our body through one test after another, trying to find out what has gone wrong, we are forced to remember that ‘Health is indeed Wealth’. 

In earlier days, life was very simple. People worked for a stipulated time, often walked everywhere, ate more homemade food, did household chores, and enjoyed a healthy balance in life. 

Now people have cars and bikes to commute, so they walk less. With the demand for more working hours, people are awake till late nights and indulge in more junk food than home-cooked food. Modern equipment at home has reduced the labour work and increased dependency on this equipment.  People don’t have enough time to exercise or even get enough sunlight. Basically, nowadays people are living very unhealthy lifestyles. 

Unhealthy living conditions have increased the contraction of people to various diseases like obesity, diabetes, heart attacks, hypertension, etc. This has alarming implications in our near future. So it is very important to focus on our health as much as we focus on our work. Moderation in food habits, daily exercise, and balanced work-life can surely make a big difference to our health and body. When a person stays mentally and physically fit, his action and decision are more practical and logical and hence he is more successful in life. Furthermore, good health has a direct impact on our personality. 

Importance of Good Health:

A healthy body has all the major components that help in the proper functioning of the body. The essential component is the state of physical health. Your life term extends when you maintain good physical fitness. If you are committed to exercising with a sensible diet, then you can develop a sense of well-being and can even prevent yourself from chronic illness, disability, and premature death. Some of the benefits of increased physical activity are as follows.

Importance of Good Health :

A healthy body has all the major components that help in the proper functioning of the body. The essential component is the state of physical health. Your life term extends when you maintain good physical fitness. If you are committed to exercising with a sensible diet, then you can develop a sense of well-being and can even prevent yourself from chronic illness, disability, and premature death. Some of the benefits of increased physical activity are as follows.

Conclusion:

Health is Wealth because if we are not healthy then all our wealth, fame and power can bring no enjoyment. Keeping fit and healthy is indeed not an option but a necessity

Main reason behind petrol price hike, the Union minister of petroleum and natural gas Dharmendra Pradhan explanis…”There has been a jump in crude oil prices in the international market.One of the main reasons behind the rise in fuel prices in india is that we have to import 80% of the oil we consume”.The bonds, he says, is what the current government has to repay along with its interest, leading to and increase in petrol rates.

He had earlier said that it is up to the GST(Goods and Service Tax) Council to decide whether the fuel should be brought under the Goods and Services Tax, which, many believe, would substantially bring down the prices.

Currently, the central and state taxes make up for 60% of the retail selling price of petrol and over 54% of diesel. Centre levies ₹32.90 per litre of excise duty on petrol and 31.80 a litre on diesel.In view of this, Prime Minister Narendra Modi-led government is under increasing pressure to cut high taxes on fuel.

-MEETALI SONI

WHAT IS REINFORCEMENT LEARNING?

Reinforcement learning (RL) is an area of machine learning concerned with how intelligent agents ought to take actions in an environment to maximize the notion of cumulative reward. It is one of three basic machine learning paradigms, alongside supervised learning and unsupervised learning.

The environment is typically stated in the form of a Markov decision process (MDP) because many reinforcement learning algorithms for this context use dynamic programming techniques. The main difference between the classical dynamic programming methods and reinforcement learning algorithms is that the latter does not assume knowledge of an exact mathematical model of the MDP and they target large MDPs where exact methods become infeasible.

HOW DOES RL WORK? In reinforcement learning, developers devise a method of rewarding desired behaviors and punishing negative behavior. This method assigns positive values to the desired actions to encourage the agent and negative values to undesired behaviors.This programs the agent to seek long-term and maximum overall reward to achieve an optimal solution.

These long-term goals help prevent the agent from stalling on lesser goals. With time, the agent learns to avoid the negative and seek the positive. This learning method has been adopted in artificial intelligence (AI) as a way of directing unsupervised machine learning through rewards and penalties.

APPLICATIONS AND EXAMPLES OF RL

Gaming is likely the most common usage field for reinforcement learning. It is capable of achieving superhuman performance in numerous games. A common example involves the game Pac-Man.

In robotics, reinforcement learning has found its way into limited tests.Itis also used in operations research, information theory, game theory, control theory, simulation-based optimization, multiagent systems, swarm intelligence, statistics, and genetic algorithms.

CHALLENGES OF APPLYING RL

Reinforcement learning, while high in potential, can be difficult to deploy and remains limited in its application. One of the barriers to the deployment of this type of machine learning is its reliance on an exploration of the environment.

For example, if you were to deploy a robot that was reliant on reinforcement learning to navigate a complex physical environment, it will seek new states and take different actions as it moves. It is difficult to consistently take the best actions in a real-world environment, however, because of how frequently the environment changes.

The time required to ensure the learning is done properly through this method can limit its usefulness and be intensive on computing resources. As the training environment grows more complex, so too do demands on time and compute resources.

Supervised learning can deliver faster, more efficient results than reinforcement learning to companies if the proper amount of data is available, as it can be employed with fewer resources.

COMMON RL ALGORITHMS

Rather than referring to a specific algorithm, the field of reinforcement learning is made up of several algorithms that take somewhat different approaches. The differences are mainly due to their strategies for exploring their environments.

1. State-action-reward-state-action (SARSA): –

This reinforcement learning algorithm starts by giving the agent what’s known as a policy. The policy is essentially a probability that tells it the odds of certain actions resulting in rewards, or beneficial states.

1. Q-learning: – This approach to reinforcement learning takes the opposite approach. The agent receives no policy, meaning its exploration of its environment is more self-directed.
2. Deep Q-Networks: – These algorithms utilize neural networks in addition to reinforcement learning techniques. They utilize the self-directed environment exploration of reinforcement learning. Future actions are based on a random sample of past beneficial actions learned by the neural network.

HOW IS RL DIFFERENT FROM SUPERVISED AND UNSUPERVISED LEARNING?

1. Supervised learning: – In supervised learning, algorithms train on a body of labeled data. Supervised learning algorithms can only learn attributes that are specified in the data set. Common applications of supervised learning are image recognition models. These models receive a set of labeled images and learn to distinguish common attributes of predefined forms.
2. Unsupervised learning: – In unsupervised learning, developers turn algorithms loose on fully unlabelled data. The algorithm learns by cataloging its observations about data features without being told what to look for.

Semi-supervised learning: –This method takes a middle-ground approach. Developers enter a relatively small set of labeled training data, as well as a larger.

1. corpus of unlabelled data. The algorithm is then instructed to extrapolate what it learns from the labeled data to the unlabelled data and draw conclusions from the set as a whole.
2. Reinforcement learning: – This takes a different approach altogether. It situates an agent in an environment with clear parameters defining beneficial activity and nonbeneficial activity and an overarching endgame to reach. It is similar in some ways to supervised learning in that developers must give algorithms specified goals and define rewards and punishments.

TYPES OF REINFORCEMENT LEARNING

Two kinds of reinforcement learning methods are:

1. Positive:

It is defined as an event, that occurs because of specific behavior. It increases the strength and the frequency of the behavior and impacts positively on the action taken by the agent.

This type of Reinforcement helps you to maximize performance and sustain change for a more extended period. However, too much Reinforcement may lead to over-optimization of the state, which can affect the results

1. Negative:

Negative Reinforcement is defined as the strengthening of behavior that occurs because of a negative condition that should have been stopped or avoided. It helps you to define the minimum stand of performance. However, the drawback of this method is that it provides enough to meet up the minimum behavior.

WHEN TO USE RL?

Here are prime reasons for using Reinforcement Learning:

1. It helps you to find which situation needs an action
2. Helps you to discover which action yields the highest reward over a longer period.

WHEN NOT TO USE RL?

1. You can’t apply the reinforcement learning model in all the situation. Here are some conditions when you should not use the reinforcement learning model.
2. When you have enough data to solve the problem with a supervised learning method

CONCLUSION Reinforcement Learning addresses the problem of learning control strategies for autonomous agents with least or no data. RL algorithms are powerful in machine learning as collecting and labeling a large set of sample patterns costs more than data itself. The key distinguishing factor of reinforcement learning is how the agent is trained. Instead of inspecting the data provided, the model interacts with the environment, seeking ways to maximize the reward. In the case of deep reinforcement learning, a neural network is in charge of storing the experiences and thus improves the way the task is performed.

Ernest Hemingway, has penned out hills like white elephants without providing much background information, clear conversation markers or information on characters’ background, not even proper names. A reader is left with several loose ends and scenarios so as to explain the story. Further, I will be explaining what I could make out from the story as a reader. 

Set near the Ebro river in Spain, the american and the girl are waiting for a train. The story advances through the short exchanges between the two characters. From the conversation, it is almost clear that the male character is trying to persuade the female character to do something. The female character, skeptical about the situation, appears to be trying to convince herself to be as confident as her companion. The characters are seen talking about “letting the air in” (Hemingway, 230) and everything being fine afterward, taking instances and conversations as proof, it feels like the man is forcing or rather trying hard to convince the female into an abortion. The man addressing it as a small operation, ensuring the girl that they will be like before after it, and disclosing that it is the only thing that is bothering him right now are the clear queues from the man’s side that it is an abortion. The girl being unsure about the situation and the man constantly trying to convince her and her unsure mind seeking validation from the man and asking him whether the procedure will make him love her again are clear points that state it is an abortion from the womens side. Without ever using the word abortion, Hemingway plants in our mind the seeds of thoughts that grow upto the conclusion that the characters are talking about abortion or rather the man is persuading the women into one. The woman comes out through the story as if she is convinced about the situation, she drinks without a limit almost as to prove she doesn’t care about the infant inside her. She dreams about her life after the procedure and seeks validation from the man, she needs to be sure that everything will be like before. It shows how she depends on his assurance, it shows how dependent she could be on the man. 

The story, which revolves around the conversation of the two characters, is open to many interpretations. As a reader, I feel that the woman is trying to strike a balance between what she wants and what he wants, but in the end she realises her old life is precious to leave just like that. 

Cake, the word itself makes us joyful, it being a part of our life from more than an era, in our birthdays, Christmas, for saying goodbye, or celebrating our small to big success. It is always present in our small to big events
But how it started, it a story of 1883 when the country’s first cake was made.

In 1880, in a coastal town of Thalasserry, Kerala, A businessman Mambally Bapu who ships milk, tea, and bread to British troops in Egypt decided to start his bakery shop to popularise baked stuff in the local, as that time there used to be just one bakery in the country which was just for the British.

He had mastered the art of biscuit making in Burma, returned from there, and set up his small bakery shop named it Royal Biscuit Factory, he began with producing 40 varieties of biscuits, bread, and buns, the interesting thing was that he uses local toddy for fermenting until the British started importing yeast into the country.
But how the idea of the baking cake was started, it was few days before Christmas in the year 1883, Murdoch Brown a British planter who started a cinnamon plantation in Anjarakkandy got off his carriage walked straight to Bapu’s shop along with a rice plum cake in his hand which he bought from England. He asked Bapu to taste the cake and asked him if he can bake one just like this.

Bapu immediately agreed, Brown gave him a 10-minute demo of the basics of cake baking, handed him some bunch of the ingredients that included cocoa, raisins, dates, and more dry fruits, and suggested French Brandy erstwhile Mayyazhi for Christmas purpose, but Bapu had his own idea about this project. He made the mould from a blacksmith in Dharmadam, chose rich spices from farms along the Malabar coast, and using a local brew made using cashew apple, Kadalipazham, a variety of banana he gave a desi touch to it.

On 20 December 1884, Bapu presented his cake to Brown, after tasting the cake Brown was amazed and certified as one of the best cakes he ever had, and order dozen of it. His made cake become a sensation in the locals and Bapu’s business becomes a hit, after that followed year he established bakery chains in different parts of the states.

Today, it has been more than 130 years since Bapu presented The first Christmas cake of India, Thalassery becomes the Leader in the cake industry of Kerala. Till now, every year in the time of December it is a rush in their bakeries just for that Christmas cake with the same recipes, not just local but from USA and UAE, they get huge orders every year.

This article foregrounds the challenges India is currently facing in bringing the level of air quality to a certain standard. It also discusses solutions that could be adopted to combat the national crisis.

Rising urbanisation, booming industrialisation, and associated anthropogenic activities are the prime reasons that lead to air pollutant emissions and poor air quality. It is expected that by 2030, around 50% of the global population will be residing in urban areas (Gurjar, Butler, Lawrence, et al. 2008). More than 80% of population in urban areas is exposed to emissions that exceed the standards set by World Health Organization (WHO 2016). Air pollution is one of the key global health and environmental concerns (Nagpure, Gurjar, Kumar, et al. 2016) and has been ranked among the top five global risk factors of mortality by the Health Effects Institute (HEI 2019). According to HEI’s report, particulate matter (PM) pollution was considered the third important cause of death in 2017 and this rate was found to be highest in India. Air pollution was considered to cause over 1.1 million premature deaths in 2017 in India (HEI 2019), of which 56% was due to exposure to outdoor PM_2.5 concentration and 44% was attributed to household air pollution. As per WHO (2016), one death out of nine in 2012 was attributed to air pollution, of which around three million deaths were solely due to outdoor air pollution.

The rising trends in population growth and the consequent effects on air quality are evident in the Indian scenario. For example, the megacities of Delhi, Mumbai, and Kolkata combined holds a population exceeding 46 million (Gurjar, Ravindra, and Nagpure 2016). Over the years, there has been a massive-scale expansion in industries, population density, anthropogenic activities, and the increased use of automobiles has degraded the air quality in India (Gurjar and Lelieveld 2005). In the last few decades, the greenhouse gas (GHG) emissions and other emissions resulting from anthropogenic activities have increased drastically (Gurjar and Nagpure 2016).

As per WHO (2016) estimates, 10 out of the 20 most populated cities in the world are in India. Based on the concentrations of PM_2.5 emissions, India was ranked the fifth most polluted country by WHO (2019), in which 21 among the top 30 polluted cities were in India. The Indian cities, on average, exceeded the WHO threshold by an alarming 500%.

The consistent population growth has led to an excessive strain on the energy consumption, thereby affecting the environment and the air quality of the megacities (Gurjar and Nagpure 2016). Kumar, Khare, Harrison, et al. (2015) calculated the increase in the total energy demand for both mobile and point sources and inferred that in Delhi, the energy demand had grown by 57.16% from 2001 to 230,222 TJ in 2011. A subsequent rise in energy consumption can be expected in the coming years, with no reliable sources available for monitoring the rate of energy consumption.

The continuous degradation of ambient air quality in the urban centres of India demands effective measures to curb air pollution. Though various policy measures are being introduced by the Government of India (GoI) to reduce the vehicular and industrial emissions, the extent to which these measures are implemented is questionable. The lack of infrastructural facilities, inadequacy of financial resources to implement advanced infrastructural innovations, difficulty in relocation of the industries from the urban centres even after mandatory court decisions, and above all, the behavioural patterns among people in accepting the green solutions are some of the crucial impediments on the road to environmental protection that our country seems to be struggling to overcome today.