12/7 comments

 As a long time frequent user of Google Translate, it was clear to me like what the article states about the improvement of the program over the years. I had to take a Japanese exam back in high school many years ago and back then my Japanese teacher used nothing but Google Translate for her classes. One thing that has significantly improved is the flexibility of translation in terms of the tone and formality. When I used Google Translate to prepare for my Japanese exam in high school, I had a lot of struggle with inputting colloquial texts into the program. It came out very awkwardly phrased at most times, in contrast to formal texts that would have better translations. I find that Google Translate nowadays has improved significantly in this sense because most informal texts are being translated to their actual meanings. What I loved about the article was the variety of ways Google Translate can be used to serve purposes other than study. It shows how a translation program can be beneficial to many parties in different working industries. I would love to see how the current applications of AI can be used to create these benefits. However, I do feel that translation is something that cannot be entirely trusted by AI due to certain texts and phrases having hidden/deeper meaning to it. Therefore I think there should be a limit to how much importance should be given to AI.

12/7 Comments

         I'd personally never found the improvements of Google Translate to be as dramatic as described in the article, but I do find it both cool and promising as time goes on. Machine learning and AI improvement improve at such a rapid pace that I can't help but wonder how auto-translations will be in ten years or by the end of the century. Of course, since other technology will also be improving at such a rate, I don't think it'd be farfetched to imagine a world where holograms are able to translate in real-time. Being able to cross the language barrier in an efficient and accurate manner can be so helpful in both formal and non-formal situations such as at a G20 gathering or just some friends from around the globe. While it'll obviously take quite a bit of time, I'm excited to see where auto-translations will take us. I think it'd also be cool to see auto-translations to serve as proofreading or maybe translators will find themselves looking over and fixing a translation made via Google Translate. I do think it'd take some fun out of the process, but it's an interesting possibility. I can't wait to see the various applications of auto-translation in the coming future.

12/7 Comments

 

It’s cool reading about the great strides that Google Translate has made, something I’ve definitely experienced for myself – it honestly doesn’t feel like it’s been all that long since it returned nothing but unintelligible gibberish, but nowadays, as demonstrated in the article, it seems to usually do the job well enough. I do wonder, however, if this may not be a drawback in some ways; if the output is just nonsense, it’s easy to tell, but if it seems to mostly make sense, it becomes much harder to tell when machine translation makes mistakes, as it still sometimes does.

It’s also interesting to see that the method that has seen most success being machine learning, as opposed to (presumably) attempting to approach the problem by with a more manual parsing of the syntax and lexicon of the languages in question. It reminds me of how learning a native language, as one does in their childhood through immersion, seems to come to us more naturally than learning a second language, where we’d typically study the grammar and vocabulary in a much more structured manner. In the same way, though, it might be harder to fix mistakes that are appearing consistently in a machine-learning based system, since it seems like we wouldn’t understand how it’s coming up with its answerse in nearly the same level of detail.

12/7 - Comments

 This was a very fun article to read. I am currently taking an A.I. course so this crossover was very welcomed. Although A.I. is not all machine learning, Google certainly enjoys using this to enhance their products. Google Maps could very well been a standard A.I. algorithm to find the shortest path to your destination, but Google decided to use machine learning to update routes, and predict traffic making what seem like longer routes more desirable. Of course as mentioned in the article, Google Translate uses machine learning to render natural sentences. As we read before, creating a symbolic A.I. to break down the rules of a language to another not only takes a long time, it is also incredibly difficult, that's why machine learning is so important for translation. The mentioned near the end of the section we had to read about the A.I. being able to "understand" well enough to engage in plausible conversation. I think that this is already making good progress. A.I. Dungeon is a game where you can type a prompt, and the A.I. will make a story for you. As it goes on you can add more story elements, and speak with other characters in the story. The technology, although a little crusty, is incredibly impressive and has only gotten better with each update. This ability to create stories and conversations is incredibly important when translating phrases and it will only improve.

12/7 Reading comments

 I must first admit that I did not know before reading this article that Google Translate had a revamp version where it started utilizing AI feature, so I didn't know that it was significantly improved some time ago. I would admit that it did a good job translating the example passage. For a person that uses Google Translate daily, that's very interesting knowing how far it has come. It was kind of a rule that you would never put a whole sentence or paragraph but only a word one by one, but who knows how advance its machine learning ability will become in the future. Maybe it will be able to translate some modern slangs that it picks up from its search engine service when a word becomes viral that people all around the world look up for it. Maybe it will become good enough that it can translate technical articles that does not have any nuances that require humans to understand. 

12/7 Reading

Machine learning, a subfield of artificial intelligence, only recently became increasingly prevalent due to availability of things such as GPUs and publicly available large data sets. It really is quite remarkable how the field of machine learning has suddenly become the forefront of computer science research and development. What's especially interesting with regards to applying machine learning to things such as image recognition or translation in this case, is that training a neural network to do so doesn't actually require a whole lot of human intervention. You set up a neural network architecture, give it example translations, and let it learn by itself how to differentiate between good/bad, right/wrong translations. Machine learning is essentially programmers telling computers how to learn without needing to describe what to learn. You can give a computer a whole bunch of examples of translations between certain languages, and given enough training time and a relatively solid architecture, the computer will learn by itself to distinguish between certain features (i.e. grammar structures, punctuation, etc.) Though it will most definitely require a large amount of time and programming ingenuity to program and train a neural network capable of perfect (or more realistically, near-perfect) translations, I do believe that it possible to do so in the (near?) future.

12/7 Reading

 At first, I was sure that the first example Lewis-Krausdec provided was the Google Translate version, since it seemed to have similar language and was fairly succinct. Then after reading the second version, I realized it had way too many words that were unnatural and repetitive, and that the first translation was short because the translator was able to cut the unnecessary and summarize the reading. It's so interesting seeing how far Google Translate has come, and while it is still not as natural as regular conversations, I do admit that the translations are much smoother than they were a year or two ago. It'll be fascinating to see how Google Translate further continues to build on its AI enhancements, and maybe we'll get closer to being able to translate a conversation naturally. 

12/7 (C)

Hopefully, Google will be able to perfect its translation service at some point in the near future. The anecdote about German-Arabic translations at the very beginning I think is a good example of why this would be a useful development, since refugees and immigrants who don't speak the language of the country they're moving to would benefit  immensely in many ways from the existence of such a ubiquitous translator. For one, it could help them navigate the legal system of their new state without needing the assistance of another who may in some cases not be very sympathetic to their plight (this may become more important soon as xenophobia seems to be on the rise around the world). A perfect digital translator may also help remove the language barrier between individuals, which may help reduce overall xenophobia and general fears of "aliens" and "outsiders"; It becomes much easier to see that others aren't so different from yourself when you're able to associate some kind of meaning with the sounds coming out of their mouths. I think we should probably make extra efforts to see each other as human, especially these days. So, hopefully Google will be able to get Translate to a near perfect state at some point in the near future, as it would do a lot of good I think!

(Sorry this was rambly >_<)

12/7 Reading Comments

 I found this article by Lewis-Krausdec very interesting, as I also sometimes use Google Translate to translate words that I do not know. Compared to Google Translate that I used to use in language classes in high school, I also feel that the website has become more advanced. Obviously it cannot be relied on fully, as it does not understand context or writing style, but Google Translate has become more and more accurate over the years. I thought that it was very interesting to learn the development of the technology, and how Google decided to reorganize itself around artificial intelligence. As technology (not exclusive to Google Translate) starts to develop dependence on artificial intelligence, machine learning technologies are becoming more accurate and helpful to daily life. Even Google Translate has become increasingly accurate as the technological development of the software furthers. I was impressed with the example that Lewis-Krausdec used of Samsung’s new ultrasound devices, as the company’s medical-imagining subsidiary announced the development of these devices that could detect breast cancer. As artificial intelligence’s development furthers, humans can increasingly rely on technology to carry out basic daily functions, and even more complex, dangerous jobs can be more accurately completed by machines. 

12/6 Reading

This reading was interesting cuz it reminded me how fortunate we are to have services such as Google Maps and Google Translate. "Machine learning" technologies are fairly recent, but I cannot imagine what life would be like without them. I didn't know google translate translated 140 billion words per day. That's a lottttttt. It's interesting that Google translate first started due to the refugee crisis. I enjoyed how they showed the difference between the old translate system and AI-rendered version. "Uno no es lo que es por lo que escribe, sino por lo que ha leído." The AI translated it so well compared to the old translate system. I was not sure how I felt about the talk about Google Maps and having a machine know everything about trends, and how much ppl typically spend for children, etc. It seems kind of excessive at that point, but I guess people will try that new technology if it becomes available. I will not be a fan, personally.

12/7 Reading Questions

 It's not really talked about in the article, but I'm interested what everything thinks about how the future of AI relates to user privacy, as it's something that I often think about, and specifically in relation to this sentence, "...a truly intelligent Maps could also conceivably know all sorts of things a close friend wouldn't... If an intelligent machine were able to discern some intricate if murky regularity in data about what we have done in the past, it might be able to extrapolate about our subsequent desire, even if we don't entirely know them ourselves." Will the inevitable loss of individual privacy through reliance on "free" services like Maps, Translate, Google, etc. coupled by their enormous amounts of data collection be overshadowed by the convenience and benefits to access of ever-improving "artificial intelligence" systems, like those with improvements as remarkable as Translate detailed in the article?

12/7 Reading Comments - Sarah Watanabe

     This article, combined with the one we read last week, was very interesting to think about the future of automated translations, especially for more routine translations. Though I hadn't been conscious of the day on which it occurred, I did definitely notice an upgrade in Google Translate's abilities to translate more extensive phrases and make it sound much more natural than the more "robotic" translations of the past. I did find the article's discussion regarding what is considered AI to be interesting because though I had never really thought about it, most things that were once AI seem to be losing that title as its function is becoming more generally available to the public and routine. Another aspect of this new Google Translate is its ability to learn and this raised the question for me about Japanese translations. Because Japanese has hiragana, katakana, and kanji, it made me wonder how well Google Translate can produce the same result if the same sentence is typed with hiragana, katakana, or kanji. In the same vein, if a person makes a mistake in the kanji they select to describe something, a human translator, though they may lead to a mistranslation, can determine this mistake and determine the meaning from context, whereas it is more probable that a machine would have more trouble with this error. But on the flip side, as we discussed during the mistakes in translation homework, humans make errors too, so if these translation technologies become very advanced, they may eventually be able to proofread for human translators as a way to prevent some of these errors. 

12/4 Reading

 I had never considered how google translate works before. As it was before I believe it worked by cataloging the meanings of each word and trying to understand whole sentences based on the dictionary definitions of each word. Of course, coinciding with the robotic understandings of grammar. But with the change of Translate, using AI or referencing other translations as I have heard it mentioned, has done much better at producing machine translations than it ever has been. I wonder, with this new system, how it responds to rapidly changing elements of language. I think it would be interesting if it could recognize new developments in language, for example; new jargon, slang, colloquial phrases. Right now Translate seems to work similarly to a child repeating things it has heard rather than truly understanding and producing that knowledge. As machine learning continues to develop, I wonder how much Translate and other AIs will be able to understand. However, I think there is still a lot to consider with the nuances apparent in language, and I am still skeptical of how well Translate would be able to recognize them. I might just be worried though that it will try to put me out of a job. But hopefully, I have a while until that might happen.