AI News, Could machine learning save this sea cow?

Could machine learning save this sea cow?

(It’s a little like the image recognition that lets you search Google Photos for shots of particular dog species, or sunsets, or whatnot—but much more specialized for this scientific task.) The results are encouraging: an early version of their detector could find 80% of the sea cows they’d found manually in images, and they expect to improve performance over time.

Eventually if they’re able to track these threatened populations on a large scale, conservationists have a much better shot at knowing how they’re impacted by human activities, and where it’s most urgent we protect their habitats.

Could machine learning save this sea cow?

(It’s a little like the image recognition that lets you search Google Photos for shots of particular dog species, or sunsets, or whatnot—but much more specialized for this scientific task.) The results are encouraging: an early version of their detector could find 80% of the sea cows they’d found manually in images, and they expect to improve performance over time.

Eventually if they’re able to track these threatened populations on a large scale, conservationists have a much better shot at knowing how they’re impacted by human activities, and where it’s most urgent we protect their habitats.

Here's how scientists use Google's machine learning technology to protect endangered sea cows

Can Google's machine learning technology help sea mammals such as sea cows that are under the threat of extinction?

Amanda Hodgson of Murdoch University and her team, along with computer scientist at Queensland University of Technology Frederic Maire, have created a detector using Google's TensorFlow platform that can automatically find sea cows in aerial photos of oceans taken from drones.

'Eventually if they're able to track these threatened populations on a large scale, conservationists have a much better shot at knowing how they're impacted by human activities, and where it's most urgent we protect their habitats.


The dugong is largely dependent on seagrass communities for subsistence and is thus restricted to the coastal habitats which support seagrass meadows, with the largest dugong concentrations typically occurring in wide, shallow, protected areas such as bays, mangrove channels, the waters of large inshore islands and inter-reefal waters.

The dugong is easily distinguished from the manatees by its fluked, dolphin-like tail, but also possesses a unique skull and teeth.

Despite being legally protected in many countries, the main causes of population decline remain anthropogenic and include fishing-related fatalities, habitat degradation and hunting.

With its long lifespan of 70 years or more, and slow rate of reproduction, the dugong is especially vulnerable to extinction.

Dugong dugon is the only extant species of the family Dugongidae, and one of only four extant species of the Sirenia order, the others forming the manatee family.[6]

The two extant families of sirenians are thought to have diverged in the mid-Eocene, after which the dugongs and their closest relative, the Steller's sea cow, split off from a common ancestor in the Miocene.

It has thick, smooth skin that is a pale cream colour at birth, but darkens dorsally and laterally to brownish-to-dark-grey with age.

The body is sparsely covered in short hair, a common feature among sirenians which may allow for tactile interpretation of their environment.[14]

These hairs are most developed around the mouth, which has a large horseshoe-shaped upper lip forming a highly mobile muzzle.[12]

A male's testes are not externally located, and the main difference between males and females is the location of the genital aperture in relation to the umbilicus and the anus.[18]

The lungs in a dugong are very long, extending almost as far as the kidneys, which are also highly elongated in order to cope with the saltwater environment.[12]

The female's tusks continue to grow without emerging during puberty, sometimes erupting later in life after reaching the base of the premaxilla.[12]

The full dental formula of dugongs is, meaning they have two incisors, three premolars, and three molars on each side of their upper jaw, and three incisors, one canine, three premolars, and three molars on each side of their lower jaw.[16]

Like other sirenians, the dugong experiences pachyostosis, a condition in which the ribs and other long bones are unusually solid and contain little or no marrow.

The largest individual recorded was 4.06 metres (13.32 ft) long and weighed 1,016 kilograms (2,240 lb),[12]

The full size of the former range is unknown, although it is believed that the current populations represent the historical limits of the range,[6]

Recorded numbers of dugongs are generally believed to be lower than actual numbers, due to a lack of accurate surveys.

and large numbers are also found in wide and shallow mangrove channels and around leeward sides of large inshore islands, where seagrass beds are common.[6]

although in areas where the continental shelf remains shallow dugongs have been known to travel more than 10 kilometres (6 mi) from the shore, descending to as far as 37 metres (121 ft), where deepwater seagrasses such as Halophila spinulosa are found.[6]

The eastern side of the Red Sea is home to large populations numbering in the hundreds, and similar populations are thought to exist on the western side.

Dugong populations in Madagascar are poorly studied, but due to widespread exploitation it is thought they may have severely declined, with few surviving individuals.[26][27]

Recoveries of seagrass beds along former ranges of dugongs, such as the Chilika Lake have been confirmed in recent years, rising hopes for re-colorizations of the species.[38]

Despite these efforts, numbers continue to decrease, and in 2007 it was reported that no more dugong could be found on the west coast of the island of Hainan.[42]

Nonetheless, dangerously low levels of awarenesses for conservation of marine organisms in Vietnam and Cambodia may result in increased intentional or unintentional catches, and illegal trade is a potential danger for local dugongs.[45]

The Gulf of Thailand was historically home to large number of the animals, but none have been sighted in the west of the gulf in recent years,[6]

A single individual was recorded in Amami Oshima, at the northernmost edge of the dugong's historic range, more than 40 years after the last previous recorded sighting.[56]

Dugong populations in these areas were reduced by historical hunts as payments to the Ryukyu Kingdom, before being wiped out because of large-scale illegal hunting and fishing using destructive methods such as dynamite fishing after the Second World War.

Some theorise that populations existed independently, for example that the Okinawan population were isolated members derived from the migration of a Philippine subspecies.[63]

this reef area houses a stable population of around 10,000, although the population concentration has shifted over time.

This population possibly shared ancestry with the Red Sea population, and the Mediterranean population had never been large due to geographical factors and climate changes.[70]

They have few natural predators, although animals such as crocodiles, killer whales, and sharks pose a threat to the young,[11]

Although they are social animals, they are usually solitary or found in pairs due to the inability of seagrass beds to support large populations.[13]

Visual communication is limited due to poor eyesight, and is mainly used for activities such as lekking for courtship purposes.

Mothers and calves are in almost constant physical contact, and calves have been known to reach out and touch their mothers with their flippers for reassurance.[13]

Dugongs are semi-nomadic, often travelling long distances in search of food, but staying within a certain range their entire life.[13]

Dugong movements mostly occur within a localised area of seagrass beds, and animals in the same region show individualistic patterns of movement.

In areas where there is a large tidal range, dugongs travel with the tide in order to access shallower feeding areas.

Although they are marine creatures, dugongs have been known to travel up creeks, and in one case a dugong was caught 15 kilometres (9.3 mi) up a creek near Cooktown.[11]

A wide variety of seagrass has been found in dugong stomach contents, and evidence exists they will eat algae when seagrass is scarce.[12]

The chemical structure and composition of the seagrass is important, and the grass species most often eaten are low in fibre, high in nitrogen, and easily digestible.[6]

Only certain seagrass meadows are suitable for dugong consumption, due to the dugong's highly specialised diet.

Feeding trails have been observed as deep as 33 metres (108 ft), and dugongs have been seen feeding as deep as 37 metres (121 ft).[6]

This behavior is known as cultivation grazing, and favors the rapidly growing, higher nutrient seagrasses that dugongs prefer.[75]

The way that females know how a male has reached sexual maturity is by the eruption of tusks in the male since tusks erupt in males when testosterone levels reach a high enough level.[78]

The age when a female first gives birth is disputed, with some studies placing the age between ten and seventeen years, while others place it as early as six years.[6]

Despite the longevity of the dugong, which may live for 50 years or more, females give birth only a few times during their life, and invest considerable parental care in their young.[77]

In these areas a male will try to impress the females while defending the area from other males, a practice known as lekking.[13]

The calf nurses for 14–18 months, although it begins to eat seagrasses soon after birth.[6]

In the Philippines, dugongs are thought to bring bad luck, and parts of them are used to ward against evil spirits.

In areas of Thailand it is believed that the dugong's tears form a powerful love potion, while in parts of Indonesia they are considered reincarnations of women.

Even in the best conditions a population is unlikely to increase more than 5% a year, leaving dugongs vulnerable to over-exploitation.

The last major worldwide study, made in 2002, concluded that the dugong was declining and possibly extinct in a third of its range, with unknown status in another half.[3]

Regional cooperation is important due to the widespread distribution of the animal, and in 1998 there was strong support for Southeast Asian cooperation to protect dugongs.

Kenya has passed legislation banning the hunting of dugongs and restricting trawling, but the dugong is not yet listed under Kenya's Wildlife Act for endangered species.

Most currently live in established marine parks, where boats must travel at a restricted speed and mesh net fishing is restricted.[11]

Despite being legally protected in many countries, the main causes of population decline remain anthropogenic and include hunting, habitat degradation, and fishing-related fatalities.[5]

Most issues with industrial fishing occur in deeper waters where dugong populations are low, with local fishing being the main risk in shallower waters.[6]

The use of shark nets has historically caused large numbers of deaths, and they have been eliminated in most areas and replaced with baited hooks.[11]

Modern farming practise and increased land clearing have also had an impact, and much of the coastline of dugong habitats is undergoing industrialisation, with increasing human populations.[11]

The shallow waters are often used as a source of food and income, problems exacerbated by aid used to improve fishing.

Plans exist to build a military base close to the Henoko reef, and military activity also adds the threats of noise pollution, chemical pollution, soil erosion, and exposure to depleted uranium.[6]

The military base plans have been fought in US courts by some Okinawans, whose concerns include the impact on the local environment and dugong habitats.[54][84]

It was later revealed that the government of Japan was hiding evidence of the negative effects of ship lanes and human activities on dugongs observed during surveys carried out off Henoko reef.[85]

Food shortages can be caused by many factors, such as a loss of habitat, death and decline in quality of seagrass, and a disturbance of feeding caused by human activity.

Sewage, detergents, heavy metal, hypersaline water, herbicides, and other waste products all negatively affect seagrass meadows.

Human activity such as mining, trawling, dredging, land reclamation, and boat propeller scarring also cause an increase in sedimentation which smothers seagrass and prevents light from reaching it.

Halophila ovalis—one of the dugong's preferred species of seagrass—declines rapidly due to lack of light, dying completely after 30 days.

Extreme weather such as cyclones and floods can destroy hundreds of square kilometres of seagrass meadows, as well as washing dugongs ashore.

Most measures for protection involve restricting activities such as trawling in areas containing seagrass meadows, with little to no action on pollutants originating from land.

dugongs are expensive to keep in captivity due to the long time mothers and calves spend together, and the inability to grow the seagrass that dugongs eat in an aquarium.[11]

Gracie, a captive dugong at Underwater World, Singapore, was reported to have died in 2014 at the age of 19, from complications arising from an acute digestive disorder.[95]

TensorFlow: an Open-Source AI Platform

It was originally developed by researchers and engineers working on the Google Brain Team within Google's Machine Intelligence research organization for the purposes of conducting machine learning and deep neural networks research, but the system is general enough to be applicable in a wide variety of other domains as well.

TensorFlow has helped researchers, engineers, artists, students, and many others make progress with everything fromlanguage translationtoearly detection of skin cancerandpreventing blindness in diabetics.The new version is faster and flexible.It includes new tf.keras module that provides full compatibility withKeras, another popular high-level neural networks library.

But in very near future, fully managed distributed training and prediction services such asGoogle Cloud Machine Learningwith TensorFlow may open the power of large and deep neural networks to everyone.

T2T facilitates the creation of state-of-the-art models for a wide variety of ML applications, such as translation, parsing, image captioning and more, enabling the exploration of various ideas much faster than previously possible.

The T2T library is built with familiar TensorFlow tools and defines multiple pieces needed in a deep learning system: data-sets, model architectures, optimizers, learning rate decay schemes, hyper parameters, and so on.Basically, you can pick any data-set, model, optimizer and a set of hyper parameters, and run the training to check how it performs.

It no longer makes sense to have separate tools for researchers in ML and people who are developing real products.” – says Greg Corrado, a senior research scientist at Google.

Marine Biology Australian marine biologists are using TensorFlow tofind sea cowsin tens of thousands of hi-res photos to better understand their populations, which are under threat of extinction.

The system was able to detect sea cows in tens of thousands of images, which relieved scientists from doing a longer and much harder manual work Farming An enterprising Japanese cucumber farmer trained a model with TensorFlow tosort cucumbersby size, shape, and other characteristics.

For this case, the system uses Raspberry Pi 3 as the main controller to take images of the cucumbers with a camera, and in a first phase, runs a small-scale neural network on TensorFlow to detect whether or not the image is of a cucumber.

And because they were hard coded, the rules were applied to every guest uniformly, rather than taking into account the unique values that could create the kind of a personalized experience that keeps guests coming back.

There are hundreds of signals that are pulled into the search rank model, which then the machine learning algorithm figures out how all the signals interact, to produce personalized search rankings.

Explanation of Boiler Feed Water & It's Treatment - Magic Marks

Dragonfly Education is an education company, that is building proprietary education content for higher learning in technical streams. We are enabling B.Tech ...

Loud Luxury feat. brando - Body (Official Lyric Video)

Stream Loud Luxury – Body now on Spotify: Listen or download ..

REAL Mermaid Sightings Around The World!

Check out these real mermaid sightings around the world! This top 10 list of amazing discoveries of real life mermaid bodies is absolutely intrigueing! Subscribe ...

700 Pound Snake Pulled out of Lake in North Carolina

Giant 700 pound snake caught in lake in Proctor, North Carolina. The giant man eating python was measured at 98 feet long. Police were contacted to see if ...

GIANT African Land Snail! | Baby Snails | My Pet Snail

My GIANT African Land Snail & Her Babies! My pet snails. Hey guys! Today I decided to show off my giant african land snail, Shrek. Shrek was rescued from ...

Economic Systems and Macroeconomics: Crash Course Economics #3

In which Jacob Clifford and Adriene Hill teach you about Economic Systems and Macroeconomics. So, economics is basically about choices. We'll look at some ...

FISH | Animals for children. Kids videos. Kindergarten | Preschool learning

FISH | Animal videos especially made for children. Made in the UK. Quality & educational videos. Click here to subscribe: ..

Underwater Whale Sounds - Full 60 Minute Ambient Soundscape

Download the Mp3 version at: ▻ Subscribe to my channel and be updated with my latest tracks: ..

How does a Helicopter fly ?

Please support us at so that we can add one more member to the team and will be able to release 2 educational videos/month.

Earthlings [Documentary]

"Earthlings is a 2005 American documentary film about humanity's use of other animals as pets, food, clothing, entertainment, and for scientific research.