You probably couldn't find much evidence re Krugman + Charlatan, because he's not. He's an accomplished economist who's influential and respected among the academic community. He's won two of the most prestigious awards in economics (Nobel + John Bates Clark). He's had some peculiar arguments in the past, which were dumb imo, but nothing "Charlatan"... Science is hard and economics is unique. Krugman is known for doing post-mortems and admitting when his ideas are off. I also think that being recognized by the academic community and his peer economists is a reasonable signal that he's not an idiot.
More recently he's been a bit more into hackish punditry and is prob. most famous among regular people for his somewhat abrasive nyt column. I think your idea about and API is not ideal. Is it enough to trust some randos on hacker news to identify Charlatans?
At least on the quant side, I think the typical sentiment is that most researchers aren't interested in ops / developing infrastructure / curating datasets.
I think he's saying that a quant can work for a few years and retire unreasonably rich, or keep working at a place where they are well rewarded and everything works.
Or they can take a strategy built on advantageous relationships with banks providing credit for leverage, an accurate and clean history of the markets and prior data to feed models, all run by teams who know what they're doing and who are constantly working to improve the edge the entire firm has, and try to do it all themselves after only really working in one small area.
It sounds like you aren't really interested in a rational discussion by the second half of your post, but the typical arguments (incl in the post) for are that market makers reduce inefficiencies in the market & provide liquidity that significantly reduces the bar (i.e. make trading cheaper) for retail investors (like you or me) to trade.
I think it is generally accepted that society does benefit from a modern and efficiently run market. Whether or not automated market makers contribute to this could be up for debate, I guess.
I don't do stock market trading, but even those who do that I know of, are doing so via companies such as: Robin Hood, E-Trade, Fidelity, Charles Schwab, Vanguard...
Are these "market makers" working behind the scenes to facilitate the operation of those retail facing companies? Is Black Rock buying all the real estate also good for (potential) retail investors like me? Because it's starting to feel like we're being told to cheer for those faciliting the ever-increasing wealth disparity of society.
>Are these "market makers" working behind the scenes to facilitate the operation of those retail facing companies
Yes. Brokers like Fidelity have no idea how to price things, and even when they do, they don't know know how to manage the risk. Marker makers quote at the tightest prices they can offer and you trade against them, through your broker, on or off-exchange.
Market makers are often much more efficient and automated than brokers, but have similar or lower margins as a business and take a lot more risk. There's a misguided anger directed to electronic market makers, but it's in fact brokers that've been ripping you off all along.
You probably do, indirectly through an agency agreement, for example a pension fund that manages your money. Or even whenever you just buy an ETF to invest. The costs you're indirectly paying are lower due to the newer generation of market makers that have reduced transaction costs for you.
> Is Black Rock buying all the real estate also good for (potential) retail investors like me?
Investing in real estate for years is not related to market making stocks with a holding period of 5 minutes.
I don't have a pension, or a 401k if that is what you are implying. I do have social security deducted from my paycheck. Is that money getting invested into the stock market on my behalf?
Sorry I'm not in the elite income class, I'm not directly familiar with the nuances of all these financial companies, or what they do. I understand risk. I understand lending money to pursue a risky venture. I understand time-value of money. I don't understand higher-order financial engineering except as presented in pop culture references such as wolf of wallstreet which I initially referenced, or the big short. I understand many machinations of society aren't directly visible as a "product" to the "average joe" of society but their ultimate benefit to society can usually be explained in a way I can understand, such as insurance, loans, industrial manufacturing, and such. These financial companies, as well as lobbyists, seem to just be skilled at manipulating a system and converting it into money.
Probably by your value system I am irrational, I don't chase money as an ends unto itself. I'm trying to understand Jane Street.
> I don't understand higher-order financial engineering except as presented in pop culture references such as wolf of wallstreet
The Wolf of Wall Street wasn't doing any sort of financial engineering in the real sense of the term. They were just committing fraud with pump and dump schemes. These guys had no actual quantitative or mathematical modeling abilities whatsoever that would be required for financial engineering and modeling. They were salesmen who swindled a lot of clueless people out of their money through illegal means.
You may not directly participate in capital markets but institutions around you that society relies on do. They do so to secure operating cash, loans, buy or sell insurance, etc. When people participate in capital markets they do so looking to make a profit or to purchase some utility, ideally these people have done some research about their trade before firing. Market makers compete for the right to charge you a fee (the spread) to make that transaction. You’re paying a fee to sell them risk (the risk that you’re correct with your opinion). Collectively this adds up to information exchange between all parties becoming less expensive: more participants on either side of any trade, smaller spreads, etc. Less friction.
Options MMs are more or less directly buying and selling insurance.
While you yourself might not be directly involved with this, a lot of what makes capitalism go round ultimately goes back to these large institutions swapping vast sums of money around and market makers help facilitate a lot of that action.
Business loans, your savings account, your employer’s (or contractor’s if you’re freelance) line of credit, the global currency system, the prices of commodities that get turned into the physical products that we consume, etc.
Well, that’s the idea anyways. Whether or not the snake has consumed it’s own tail is a whole different discussion, but the stated value of stuff like this is to create efficient markets with correct price/price discovery aka make sure no one is paying too much or selling for too little.
I think the equation changes when you weigh the marginal utility that market makers provide against the social opportunity cost of allocating intelligence to these firms.
You think you waited a long time? I wrote a patch to add this to GitHub when I worked there in 2011-12! :) It was rejected at the time by some folks for various “product” reasons, but I’m glad to see it finally become a thing.
I think K is usually interpreted as the # of gaussians from which the data has been assumed to be sampled. Not immediately related to # of latent variables unless you invoke kernel kmeans or something like laplacian eigenmaps/diffusion maps.
The term "labeled graph" just means a graph with each node labeled differently (but arbitrarily). It just allows for reasoning & enumerating the vertex set. It's a typical assumption to make in the context of graph isomorphism.
It doesn't relate to machine learning (which is what I assume you mean).
Isomorphism of a pair of graphs usually refers to isomorphism of their unlabelled equivalents.
Yes the concrete expression of the isomorphism would be as a mapping between the labels.
Given that the paper linked to is by Brendan McKay et al, it seems reasonable to mention that nAUTy works by finding (efficiently) all permutations of the labellings that result in an automorphism of the graph.
That's really cool. These days similar strategies based on graph coarsening and vertex ordering are really popular for improving the sparsity pattern of preconditioners- e.g. incomplete lu decompositions for iterative linear solvers.
yeah, there's so much room to extend this work in terms of graph coarsening / sparsifying, especially for progressive / approximate approaches
there's lots of overlap with entropy binning as well and likely some sort of combo for win -- figure out "max frequency" with some sort of FFT / wavelet like meta technique or somehow bin edges based on entropy contribution, etc.
or, lots of pathways to make coontinuous / numerical in JPEG like way
basically JPEGs for graphs -- with big impacts in AI / ML computes
I got Micro Snitch [1] as part of a bundle with Little Snitch years ago and have just had it running for cases like this. I'm fortunate to not have run into this issue, but I like the peace of mind of knowing exactly if I do.
I'm less familiar with poisoning, but at least for test-time robustness, the current benchmark for image classifiers is AutoAttack [0,1]. It's an ensemble of adaptive & parameter-free gradient-based white-box and gradient-free black-box attacks. Submitted academic work is typically considered incomplete without an evaluation on AA (and sometimes deepfool [2]). It's good to see that both are included in ART.
A new subfield of adversarial ML that considers similar challenges to adversarial NLP: topological attacks on graphs for attacking graph/node classifiers.
Both problems (NLP & graph robustness) are made much more challenging compared to adversarial robustness/attacks on image classifiers due to their combinatorial nature.
For graphs, canonical notions of robustness wrt classes of perturbations defined based on lp norms aren't so great (e.g. consider perturbing a barbell graph by removing a bridge edge- huge topological perturbation, but tiny lp perturbation!)
I think investigating robustness for graph classifiers should also help robustness for practical nlp systems and visa-versa. For example, is there any work that investigates robustness of nlp systems, but considers classes of perturbations defined on the space of ASTs?
More recently he's been a bit more into hackish punditry and is prob. most famous among regular people for his somewhat abrasive nyt column. I think your idea about and API is not ideal. Is it enough to trust some randos on hacker news to identify Charlatans?