From the paper's method section, a bit more about which type of ML algo was used:
An RF machine-learning model was developed to predict lithium concentrations in Smackover Formation brines throughout southern Arkansas. The model was developed by (i) assigning explanatory variables to brine samples collected at wells, (ii) tuning the RF model to make predictions at wells and assess model performance, (iii) mapping spatially continuous predictions of lithium concentrations across the Reynolds oolite unit of the Smackover Formation in southern Arkansas, and (iv) inspecting the model for explanatory variable importance and influence. Initial model tuning used the tidymodels framework (52) in R (53) to test XGBoost, K-nearest neighbors, and RF algorithms; RF models consistently had higher accuracy and lower bias, so they were used to train the final model and predict lithium.
Explanatory variables used to tune the RF model included geologic, geochemical, and temperature information for Jurassic and Cretaceous units. The geologic framework of the model domain is expected to influence brine chemistry both spatially and with depth. Explanatory variables used to train the RF model must be mapped across the model domain to create spatially continuous predictions of lithium. Thus, spatially continuous subsurface geologic information is key, although these digital resources are often difficult to acquire.
Interesting to me that RF performed better the XGBoost, would have expected at least a similar outcome if tuned correctly.
RF is a heavy hitter when it comes to tabular data. XGBoost is good as well, but more often than not needs and autotuner to really unlock it (e.g pycaret).
Extracting lithium from brine is cleaner than e.g. extraction from spodumene ore. Also direct lithium extraction from brine is faster, cleaner, smaller footprint, lower energy consumption.
Uranium is extracted by mining the surface. I don't know ore concentrations though, so maybe not much land area is needed since it is a dense energy source?
I read the article carefully, twice. Doesn't have a link to any original paper, of course. And I can't find the answer to my question... did they, you know, validate the model? Did they actually take some samples at new locations and compare it to what the model says?
Or are they literally just announcing that "Hey, we told the computer to tell us something, so it told us something"? Yes, that is how it works. The computer will always tell you something if you make it tell you something. That isn't the hard part. The hard part is getting it to tell you things that correspond to reality.
In the absence of validation, this means very little, especially in an environment where the USGS is fairly incentivized to loudly announce to the world that we've totes got plenty of lithium, my fellow countries, any effort to keep lithium away from us is just a waste of time, look at us just rolling in lithium over here.
Or, maybe they did do the validation, and it's just the reporting that doesn't consider that an important aspect of the story. Somewhere between funding and press release someone's lost the trail but I don't know who exactly.
Say Lithium becomes essentially free because we find so much of it…would that drastically lower battery costs? Is our current supply of lithium limiting production?
Is sand essentially free because we have beaches and deserts full of it? It can be used to make concrete, a valuable material? (Don't forget the shipping, storage, and refining costs)
The major limiting factor of lithium is not really availability so much as the cost of extraction. China is the leader in this field, not so much because of abundance or stellar technology, but out of a willingness to completely ignore environmental externalities (including those of the power generation involved in the whole process). As a result the price of Chinese lithium is low enough that it would be essentially impossible to compete with them unless a country had similar... "advantages"... or some new and impressive technology.
In the US environmental regulations, the cost of producing power, labor costs, would all drive up the price of the end product in a way that makes it totally noncompetitive. That's also why the US and some other countries are investing in other ways to find lithium (among other things) on seabeds, where it's hoped that extraction would be less expensive. Of course the threat to the seabed environment is a concern, which in turn might drive up prices by imposing regulation, etc etc etc.
> the price of Chinese lithium is low enough that it would be essentially impossible to compete with them...
In an export model, yes. However, given their negative externalities (including geo-political factors), importing countries may place tariffs on Chinese lithium in order to make use of other sources.
If the total embodied value of lithium in any particular product is small compared to the overall value of the product, the tariff might not represent a significant drag on the indigenous industry.
This goes double if your refining and battery production is still in China as well. If you are using the material domestically then the situation could be made more fair with tariffs, but if you're exporting that obviously won't work.
Lithium is too abundant in the world right now (as expected - we've just gotten better at discovering it).
To be honest, the energy problem is more or less a solved problem with the current technologies we have. We just need to accelerate our pace of adoption to hard-reverse on fossil fuels (except Germany). We already have large reserves of Uranium, of which only a small amount is needed to fuel a power plant. We already have lithium battery tech to store the power. We already have solar panels being mass produced and adopted to fill in the gaps. All we need is connecting the dots and making sure these resources play well with each other in symbiosis.
Great, now ask the AI to engineer a fungal genome that'll help us purify it more easily: Frack in the substrate and spores, harvest fruit bodies on the surface, profit.
I would have guessed better results in the 1am to 2am time slot, but 3am is not totally out of line. I bet the fraction of drivers at 3am that are drunk is much higher than at, say, 3pm.
Towns that make a living by ticketing people passing through.
The worst place in the world for this is Italy. Every time I go there they find some esoteric rule to ticket me for. This time in Padova, apparently I drove in an area where only locals are allowed to drive. Bunch of swindlers.
Indeed in Italy there are area (mostly historical centres) where cities limit the influx of cars to keep it liveable and walkable, therefore only residents are allowed to bring their car in.
People downvoted you to the point that your comment is grayed out and about to be hidden but there is hardly metric by which Arkansas is not in the bottom ten on a list of states.
Infant mortality rate? 3rd most deadly for babies.
Poverty rate? 7th poorest.
Homicide rate? 7th most dangerous.
Obesity rate? 3rd fattest.
Practically any map of any measurable statistic where states are colored red for "bad" and green for "good" Arkansas will be a deep, blood, red.
You are citing the US News "best states" ranking. In that ranking, California is ranked #37 overall and Arkansas is ranked #47 overall. Even your own hand picked data source supports the OP...
37th/50 isn't good. But people never clamor on about how awful California is every time it's mentioned(well, rarely). This same ranking puts states like South Dakota and Indiana ahead, which I'm sure many would object to all the same.
This is the second day in a row I've watched threads about Arkansas of all places devolve into these nasty generalities(yesterday's was about WalMart and Bentonville). I don't live in Arkansas or anything, but I think we as a community can do better than devolve into it over and over, unless the topic at hand was the problems of a state.
I'm not saying that 37 out of 50 is good. I'm saying that 47 out of 50 is bad. Your data source doesn't refute the OPs argument that Arkansas is not a great place to live -- it actually supports the OP.
People vote in good faith, I presume. Sometimes a comment’s factual basis matters less than its overall contribution to a productive and open discussion. Downvotes in this case are an example of HN’s surprisingly effective system for self-moderation working as it should. It isn’t vile enough to censor, but it also isn’t what a lot of readers come here for. It didn’t personally offend me (I didn’t vote either way), but I take occasional downvoting that I don’t fully agree with in stride, as the overall system seems to work better than most.
My guess is that the presence of lithium in the groundwater is in trace amounts if at all, while the dosing of lithium is in the domain of ~300mg. A casual search for the quantity of lithium in brine from a mine shows a max of 1400ppm for a rich mine in Chile[1] so drinking straight brine wouldn't get you anywhere near the therapeutic dose. Good question!
I am not a health researcher or anyting, but a quick googling seems to suggest its possible that it lowers risks of suicide[0] and other affective disorders, which by extension it would lower the rates of issues that can contribute to these issues I'd think.
That said, I honestly am unsure. It also is a requisite that it must be in the water in sufficient but low amounts
The formation is 7000 feet below the surface, if I understand correctly, so I don't think there would be any communication of its brine with potable groundwater.
I would like to think that if there were any interaction between theses putative deposits to the groundwater that we wouldn't have needed an ML model to find these deposits in the first place!
When the tide goes out on the AI hype there’s going to be a lot of companies currently using expensive API calls for simple classification tasks that will be quietly revamped to use a simple CNN.
ML is a toolbox of methods. Not every problem needs a transformer.
The USGS predictive model provides the first estimate of total lithium present in Smackover Formation brines in southern Arkansas, using machine learning, which is a type of artificial intelligence.
I was disappointed in that line. They could’ve mentioned it used a random forest, which is much more informative. “ML is a type of AI” isn’t even a cocktail party understanding of the topic.
Are we getting to the critical point where we declassify a bunch of stuff as AI? Used to be expert systems were considered AI. Now anything-not-an-LLM is going to stop being AI?
Isn’t it a critical component of everything currently sporting anything remotely close to a legit “AI” label? I wouldn’t call cows “one part of a broader beef ecosystem” for example. They’re fundamental to it.
I think there will be markets for many different chemistries and there's unlikely to be some total winner in the near future. Each chemistry has its own tradeoffs and use cases. Some will fade and die over time like Ni-Cad, but even that takes longer than you would expect.
It would be amazing for some low weight, low volume, high energy density, high discharge rate, high charge rate, cheaply manufactured from abundant materials, low thermal sensitivity, high thermal tolerance, low passive loss, non-explosive, high cycle count, low memory, shelf stable battery chemistry to appear, but thus far every one fails in several of the categories.
Interesting, & not necessarily in a good way. This method could well presage unprecedented numbers of attempts at eminent domain takings or other means of forcing people out of their properties.
Which government agency would use eminent domain to take land and start mining? We have historical precedence with the oil industry using various scanning methods in a similar manner, but it was the oil companies who went to the landowners to acquire the rights to extract. Then the government would buy the (usable) product from them.
National security (by identifying and processing rare earth metals and materials domestically) is vastly more important to society than a few dozen homes somewhere.
Globally price-competitive domestic electric car production is a national security concern only if we are willing to accept a rather short time horizon and a rather narrow definition of security in our analysis.
This kind of article can perhaps be understood as an attempt to turn a federal organization's sails into the prevailing political winds, so to speak, at a time when funding seems insecure. I say this as someone who strongly supports most of the survey's mission. It would be ideal if national power brokers recognized the value of water science, geology, ecology, etc, on their own terms.
> Mineral rights are automatically included as a part of the land in a property conveyance, unless and until the ownership gets separated at some point by an owner/seller.
> Since sellers of land can convey only property that they own, each sale of the land after the minerals are separated automatically includes only the land. Deeds to the land made after the first separation of the minerals will not refer to the fact that the mineral rights are not included.
> in most cases, you cannot determine whether you own the rights to the minerals under your land just by looking at your deed. Owners are sometimes surprised to find out someone else owns the rights to the minerals under their land
> U.S. laws regulating mining and mineral rights typically prohibit mineral owners from damaging or interfering with the use of any homes or other improvements on the land when extracting minerals. As a result, mineral owners do not typically attempt mineral extraction in highly populated areas. This means that if you live in a city, or an area with many houses on small plots of land, you probably won't need to worry about whether or not you own any minerals that might be under you
And of course you would not mind owners of extraction company leaving all the profits to people who got kicked out of their home. After all they should be happy just fulfilling your "national security" goal.
An RF machine-learning model was developed to predict lithium concentrations in Smackover Formation brines throughout southern Arkansas. The model was developed by (i) assigning explanatory variables to brine samples collected at wells, (ii) tuning the RF model to make predictions at wells and assess model performance, (iii) mapping spatially continuous predictions of lithium concentrations across the Reynolds oolite unit of the Smackover Formation in southern Arkansas, and (iv) inspecting the model for explanatory variable importance and influence. Initial model tuning used the tidymodels framework (52) in R (53) to test XGBoost, K-nearest neighbors, and RF algorithms; RF models consistently had higher accuracy and lower bias, so they were used to train the final model and predict lithium.
Explanatory variables used to tune the RF model included geologic, geochemical, and temperature information for Jurassic and Cretaceous units. The geologic framework of the model domain is expected to influence brine chemistry both spatially and with depth. Explanatory variables used to train the RF model must be mapped across the model domain to create spatially continuous predictions of lithium. Thus, spatially continuous subsurface geologic information is key, although these digital resources are often difficult to acquire.
Interesting to me that RF performed better the XGBoost, would have expected at least a similar outcome if tuned correctly.
reply