【推荐1】On Tuesday, Virgin Atlantic flew a large passenger jet from London to New York using 100% “Sustainable Aviation（航空） Fuel（SAF）”. The flight was meant to show that it’s possible to fly using cleaner fuels.

The fuel used on the flight was mainly made from used cooking oils and animal fats. A small part of the fuel was made from corn waste. Virgin Atlantic says that using SAF cuts the flight’s pollution by 70%. SAF still pollutes when it’s burned, just like regular jet fuel. However, the difference is in how the fuels are made.

SAF is made from plants （and related animal products） that once absorbed carbon dioxide （CO₂） from the air. When SAF burns, it just releases this same CO₂ again. That’s different from jet fuel, which is made from oil exploited from underground, releasing CO₂ that was deeply buried.

SAF may sound great, but it still has many problems. For one thing, SAF costs five times as much as regular jet fuel. That helps explain why only one-tenth of 1% of the fuel airlines currently use is SAF. Virgin Atlantic is hoping that its flight will encourage more companies to produce SAF and that this will bring the price down.

But even if the price of SAF drops, critics say there are still big problems with it. They say it’s easy to make small amounts of SAF out of plant waste. But to make as much SAF as the airlines really need would require farmers to grow plants for fuel instead of for eating. This could also lead to more forests being cut down for farmland.

Airlines like SAF because it can be used now in existing planes with no changes. They hope it will help quickly reduce airplane pollution until non-polluting fuels are developed.

Governments seem to agree that SAF is a step in the right direction. Both the United States and the European Union have set targets that will sharply increase the use of SAF in coming years.

【推荐2】My cat, Momo, will meow (喵喵叫) when she is hungry and when she is full, or sometimes for what appears to be no real reason at all.

The moment I downloaded MeowTalk Cat Translator, a mobile app that could change Momo’s meows into plain English, she shut up right away. For two days I failed to appeal a sound. On Day 3, out of hopelessness, I decided to pick her up while she was eating. Momo screamed. The app gave a translation: “I’m happy!”I was doubtful. But MeowTalk provided a more reasonable translation about a week later, when I returned from a four-day trip. Upon seeing me, Momo meowed. “Nice to see you,” the app translated. Then: “Let me rest.”

MeowTalk is the product of a growing interest in joining additional intelligence to explain animal communication. Machine-learning systems, which are able to extract patterns from large data sets, can tell the difference between the sounds that cats make when they are happy and those that they give off when they are in pain. “We’re trying to understand what cats are saying and give them a voice,” said Javier Sanchez, a founder of MeowTalk. “We want to use this to help people build better and stronger relationships with their cats.”

MeowTalk uses algorithms (演算法) to identify cat sounds made in a variety of contexts. The app analyzes cat sounds in real-time, to know the intention of cats. It then displays a conversational, plain English “translation” of whatever intention it discovers, such as Momo’s annoyed “Let me rest”. MeowTalk uses the sounds it collects to improve its algorithms and its performance.

The world can be a lonely place, especially so in the last few years. Finding new ways to connect with other creatures can be a much needed comfort. Personally, I would pay at least two figures for an app that could help me know whether my dog truly needs to go outside or just wants to see if the neighbor has put bread out for birds.

【推荐3】From sweeping landscapes to amazing buildings, Taiwan is an island that many tourists visit frequently. That’s why it can seem odd that one of its most famous landmarks is two humble mailboxes. Located in the Zhongshan District in the city of Taipei, these unique mailboxes have caught the public’s imagination thanks to their appearance. Unlike most mailboxes, these two are severely bent.

Due to its location, Taiwan tends to be struck by typhoons, as it is usually hit by three or four windstorms a year. In 2015, a particularly bad one known as Typhoon Soudelor struck the island. During the storm, a billboard got loose due to the heavy winds and hit the mailboxes. While it bent their posts, it didn’t tear them down. After the storm was over, people came across the newly leaning mailboxes and fell in love with them. Not only did they look adorable, but they were also a symbol of resilience(坚韧) and survival.

Social media quickly caught wind of these new landmarks and people began visiting them. It wasn’t long before they started staging fun photos with the new landmarks, using forced perspective—similar to the way tourists pose in Italy with the Leaning Tower of Pisa. This led authorities to accept the symbolism of the mailboxes willingly, and instead of repairing them, they added a board telling their story in Chinese, English, and Japanese.

The board reads: “On August 8th, 2015, Typhoon Soudelor struck Taiwan. We were hit in the heads by a signboard as a result and couldn’t help but lean over from the pain. It was so painful in fact—we can’t stand up straight anymore. But we persevered: we didn’t fall over after all. We have to take life as it comes and in a way, I think we look more artsy and sophisticated this way! Do we inspire you? Take a picture with us to remind yourselves that we are all survivors and can persevere no matter what.”

The best part of this Taipei attraction is that the mailboxes are still operational. The letters in the boxes are even marked with a special stamp that shows the famous mailboxes.

【推荐1】Researchers at the Allen Institute for AI in Seattle released new technology that was designed to make moral judgments. They called it Delphi. Anyone could visit the Delphi website and ask it a moral question.

Psychologist Joseph Austerweil tested the technology using a few simple circumstances.When he asked if it was right to kill one person to save 100 others, it said he should. Feeling quite pleased, he then asked if he should kill one person to save 101 others. This time, Delphi said he should not. Though some users such as Austerweil found Delphi was sometimes frustrating and disturbing, most agreed that the system was surprisingly wise. Many technologists believed Delphi is an effort to address what some see as a major problem in moder AI systems: They can be as flawed as the people who create them.

Facial recognition systems and digital assistants occasionally show prejudice against women and people of color. Social networks like Facebook fail to control hate speech, despite wide use of AI. AI systems used by courts make sentencing (判刑) recommendations that can often seem illogical. Many computer scientists are working to address that issue. And the creators of Delphi hope to build a moral framework that could he used in any online service,robot or vehicle.

Delphi works like a neural(神经的)network that learns skills by analyzing large amounts of data. After gathering millions of everyday circumstances from websites and other sources,the Allen Institute asked workers on an online service-everyday people paid to do digital work at companies like Amazon-to identify each one as right or wrong. Then they fed the data-more than 1.7 million moral judgments by real live humans-into Delphi.

When the Allen Institute released Delphi, it described the system as a computational model for moral judgments. But after many complained about the obyious limitations of the system,the researchers adjusted the website. It now calls Delphi “a research model designed to copy people’s moral judgments.” It also comes with an announcement: “Model outputs should not be used for advice for humans, and could be potentially offensive or harmful.”

【推荐2】The Impact of AI on Smart Homes

Artificial Intelligence (AI) is surely changing the way we live and work, and smart homes are no exception. With AI, smart homes will become even more automated and responsive.     1    

Smart homes can learn about the people living in it. With special sensors and AI, the home can see which rooms people are in. Then the home can change things like the temperature, lights and other things in that room to make people comfortable.     2     For example, when people enter a room, the lights will automatically turn on.

Smart homes can save energy and help the environment. With AI, the home can sense when people are not using things like lights. Then the lights will turn off by themselves to conserve power.     3     Smart thermostats (恒温器) and window shades work together to keep the home cool without using much air conditioning.

AI has improved home safety a lot. It uses motion sensors and cameras around the home to watch out for things automatically. Security systems with AI can tell if something strange is happening. It will send alarms to owners on their phones.     4     Smoke and gas detectors use AI to detect dangers early, which makes homes much safer.

    5     Things people wear, like watches and activity trackers, will share data with safety devices around the house. Together they can watch people closer and make sure everyone is okay in the home.

All in all, with the increasing adoption of AI technology, smart homes are becoming more innovative, energy-efficient, safe, and convenient.

【推荐3】A deep neural (神经系统的) network trained to restore ancient Greek texts can do so with 72% accuracy when used by historians, suggests a Nature paper. The findings could assist with the restoration and sort of newly discovered or uncertain inscriptions (碑文) with improved speed and accuracy, advancing our understanding of ancient history.

To understand the history of ancient civilizations, historians study the inscriptions created by past individuals, written directly on materials- such as stone, pottery or metal- that have survived until today. However, many inscriptions have been damaged over the centuries. Their texts are now vague and their date of writing is uncertain. Specialists in the study of inscriptions can reconstruct missing texts, but their traditional methods are highly complex and time-consuming.

To overcome the restrictions of current epigraphic methods, Yannis Assael, Thea Sommerschield and their colleagues tested a deep neural network (named Ithaca), a type of artificial intelligence (AI) that was trained to restore, date and place ancient Greek inscriptions. The authors found that Ithaca could achieve 62% accuracy when used alone to restore damaged texts, and 72% accuracy when it is used by a historian. Additionally, Ithaca could also help to determine inscriptions’ place and date of writing; in their experiments, it attributed inscriptions to their original locations with 71% accuracy and dated them to less than 30 years from the date ranges proposed by historians.

The findings could unlock the cooperative potential between artificial intelligence and historians, and improve our understanding of human history.

Common fields such as education, healthcare, catering, and industry have long been dominated by artificial intelligence, and this technology is now finding its way into relatively niche (小众) fields to more fully exploit its talents. Archaeology (考古学) is one of its latest fields.