OAKLAND, Calif. (Reuters) - California regulators sued Cisco Systems Inc (CSCO.O) on Tuesday, accusing it of discriminating against an Indian-American employee and allowing him to be harassed by two managers because he was from a lower Indian caste than them.
FILE PHOTO: The Cisco logo is seen at their booth at the Mobile World Congress in Barcelona, Spain, February 26, 2018. REUTERS/Sergio Perez
U.S. employment law does not specifically bar caste-based discrimination, but California’s Department of Fair Employment and Housing contends in the lawsuit that India’s lingering caste system is based on protected classes such as religion.
The lawsuit, filed in federal court in San Jose, does not name the alleged victim. It states he has been a principal engineer at Cisco’s San Jose headquarters since October 2015 and that he was raised at the bottom of India’s caste hierarchy as a Dalit, once called “untouchables.”
That caste hierarchy was enforced in the workplace, according to the lawsuit, which accuses the two former engineering managers of harassment.
Cisco said it did not have immediate comment, saying that it planned to issue a statement at a later time.
Like other large Silicon Valley employers, Cisco’s workforce includes thousands of Indian immigrants, most of whom were brought up as Brahmins or other high castes.
The civil rights group Equality Labs in a 2018 report cited in the lawsuit found that 67% of Dalits surveyed felt treated unfairly at their U.S. workplaces.
Some employers have begun taking action. Massachusetts-based Brandeis University last year added caste to its nondiscrimination policy in what is believed to be a landmark move in U.S. higher education.
At Cisco, the unnamed employee reported his manager to human resources in November 2016 for outing his caste to colleagues. The manager allegedly retaliated, but Cisco determined caste discrimination was not illegal and issues continued through 2018, the lawsuit states.
Cisco reassigned the employee’s duties and isolated him from colleagues, rejected a raise and opportunities that would have led to one and denied him two promotions, according to the lawsuit.
After weeks of protest in the wake of George Floyd’s murder, the country was thrown into an important dialogue about police reform. Massive protests have forced ending qualified immunity, halting military transfers to police stations, disbanding police unions, and even reconsidering police budgets to rapidly become top legislative priorities.
However, the conversation has not always remained on the topic of reform. Many bad-faith actors who would like to distract from such conversations, focused instead on looting taking place during the mass demonstrations. And although the Black Lives Matter protest leaders condemned the practice, many in the media and Congress refused to talk about anything else.
So let’s talk about looting. Through civil asset forfeiture, the police have been the biggest culprits, anyway.
The process known as civil asset forfeiture allows police officers “to lawfully seize personal property from those under suspicion of committing — or helping to commit — a crime. The threshold for this practice requires only the notion of probable cause, which is considered sufficient if there is “reasonable ground for belief of guilt.”
That is, you can have your property taken from you without ever being charged with a crime.
This process is unconstitutional, and should be abolished, along with qualified immunity, military transfers, and the rest of the reform legislation, as, while harmful to everyone’s civil liberties, disproportionately affects minority communities.
Consider a few examples of how this practice has been implemented against civilians across the United States.
Last year Missouri police officers coerced at least 39 motorists into signing $2.3 million is personal assets without convictions. In March of this year, Michigan’s attorney general charged Macomb County Prosecutor with felony embezzlement of asset forfeiture funds. January 2020 saw the DEA seize a 79-year old Pittsburg man’s life savings because it was deemed suspicious to travel with large amounts of cash. And Indiana wants to seize your $42,000 car for going 5 mph over the speeding limit.
The list goes on, but some places are particularly bad.
Between the years 2012 and 2017, Reason’s CJ Ciaramella showed the Chicago police department seized $150 million worth of personal property without conviction 23,065 times— the vast majority of those instances occurring in predominantly black neighborhoods. What’s more, in Chicago’s particular case, police were found to have taken everything from cashiers checks to Xbox controllers, and a total of nearly 6,000 vehicles—which seriously disrupts people’s daily life.
This is true looting.
Ciaramella goes on to write that the “average estimated value of a seizure was $4,553, while the median value was $1,049 about three-quarters of all seizures were cash, not property.” The state of Illinois notoriously maintains prosecutors and police that keep up to 90% of all asset forfeiture in state revenue, and if you observe all assets between 2005 and 2015, the number jumps to a net worth of $319 million.
So the police departments and federal agencies like the DEA are taking in revenue to fund their programs, and black people are being forced to provide it.
This should enrage any American who cares at all about racial injustices and property rights. Moreover, this practice deserves the seriousness not many have historically taken it.
But instead many, including our own Justice Department, defend the practice as important for crime prevention. However, civil asset forfeiture has been shown to have little to no impact on fighting crime. Instead, it provides dubious profit motives to manufacture crime, as most of the property they loot from the citizens they are supposed to serve and protect lines the department’s own pockets.
What’s more, there’s hardly any way to seek justice. About 9 in 10 cases of civil asset forfeiture never make it to court, due largely to the insurmountable costs of legal action (atop the costs individuals incurred from property seizure already).
It’s obvious that action to ban this practice — this professional looting — is needed to protect vulnerable communities and stop facilitating injustices. In our current moment, it’s more important than ever to prevent bad policing. Abolishing civil asset forfeiture takes a step in that direction.
Anthony DiMauro is a writer based in New York City. His work has appeared in the Orange County Register, Real Clear, The National Interest and elsewhere. Follow him on Twitter @AnthonyMDiMauro.
Cosmic rays may have given right-handed genetic helixes an evolutionary edge at the beginning of life’s history.
Lopsided interactions between cosmic-ray particles and chiral biomolecules may explain why all life on Earth relies exclusively on right-handed DNA and RNA helixes.
Samuel Velasco/Quanta Magazine
If you could shrink small enough to descend the genetic helix of any animal, plant, fungus, bacterium or virus on Earth as though it were a spiral staircase, you would always find yourself turning right — never left. It’s a universal trait in want of an explanation.
Chemists and biologists see no obvious reason why all known life prefers this structure. “Chiral” molecules exist in paired forms that mirror each other the way a right-handed glove matches a left-handed one. Essentially all known chemical reactions produce even mixtures of both. In principle, a DNA or RNA strand made from left-handed nucleotide bricks should work just as well as one made of right-handed bricks (although a chimera combining left and right subunits probably wouldn’t fare so well).
Yet life today uses just one of chemistry’s two available Lego sets. Many researchers believe the selection to be random: Those right-handed genetic strands just happened to pop up first, or in slightly greater numbers. But for more than a century, some have pondered whether biology’s innate handedness has deeper roots.
“This is one of the links between life on Earth and the cosmos,” wrote Louis Pasteur, one of the first scientists to recognize the asymmetry in life’s molecules, in 1860.
Now two physicists may have validated Pasteur’s instincts by connecting the unvarying twist in natural DNA with the behavior of fundamental particles. The theory, which appeared in May in The Astrophysical Journal Letters, doesn’t explain every step of how life acquired its current handedness, but it does assert that the shape of terrestrial DNA and RNA is no accident. Our spirals might all trace back to an unexpected influence from cosmic rays.
This work “points out a new chiral agent which we were not considering,” said Dimitar Sasselov, an astronomer at Harvard University and the director of the school’s Origins of Life Initiative, who was not involved in the research. “It seems to be very good.”
They started from the fact that cosmic ray showers, like DNA strands, have handedness. Physical events typically break right as often as they break left, but cosmic ray particles called pions tap into one of nature’s rare exceptions. When pions decay, the process is governed by the weak force — the only fundamental force with a known mirror asymmetry. Pions slamming into the atmosphere produce showers of particles including electrons and their heavier siblings, muons, all of which are equipped by the weak force with the same chiral magnetic orientation relative to their path. The particles bounce around as they streak through the atmosphere, Globus said, but overall they tend to keep their preferred chirality when they slam into the ground.
Earth’s first organisms — which were perhaps little more than naked barber poles of genetic material — likely came in two varieties, the researchers assumed. Some had DNA or RNA strands that curled like ours, which she and Blandford dubbed “live” molecules (chiral naming conventions differ by field), and others had mirror-reversed strands — “evil” life. With a series of toy models, the researchers calculated that the biased cosmic ray particles were ever-so-slightly more likely to knock an electron loose from a “live” helix than from an “evil” one, an event that theoretically causes mutations.
The effect would be tiny: Millions if not billions of cosmic ray strikes could be required to yield one additional free electron in a “live” strand, depending on the event’s energy. But if those electrons changed letters in the organisms’ genetic codes, those tweaks may have added up. Over perhaps a million years, Globus suggests, cosmic rays might have accelerated the evolution of our earliest ancestors, letting them out-compete their “evil” rivals. “If you don’t have mutations, you don’t evolve,” she said.
The researchers’ next task is to see if the handedness of real particles can actually cause the speedy mutation seen in their model. After they published their research, Globus approached David Deamer, a biologist and engineer at the University of California, Santa Cruz, for help. Impressed by her ideas, he suggested the simplest biological test he could think of: an off-the-shelf assay known as the Ames test that exposes a bacterial colony to a chemical to find out if the substance causes mutations. But instead of evaluating a chemical, the researchers plan to roast the microbes with beams of chiral electrons or muons.
Proof that the handedness of particles really can mutate microbes would strengthen their case that cosmic rays shoved our ancestors off the evolutionary starting block, but it still wouldn’t fully explain the uniform chirality of life on Earth. The theory doesn’t address, for example, how “live” organisms and “evil” organisms managed to materialize from a primordial smoothie containing both right- and left-handed building blocks.
“That is a very hard step,” said Jason Dworkin, a senior astrobiologist at the NASA Goddard Space Flight Center and an investigator with the Simons Collaboration on the Origins of Life, “but if this [theory] can provide a different mechanism, another Darwinian pressure, that’d be interesting.”
Even before genetic evolution enters the picture, another unknown process seems to handicap “evil” life. The simple amino acid molecules that form proteins also exist in “live” configurations favored by life and “evil” configurations that are not (although the preferred chirality for “live” amino acids is almost exclusively left-handed). Careful analysis of meteorites by Dworkin and others has found that certain “live” amino acids outnumber “evil” ones by 20% or more, a surplus they may have passed on to Earth. The excess molecules could be the lucky survivors of billions of years of exposure to circularly polarized light, a collection of beams all spiraling in the same direction that, experiments have shown, can destroy one type of amino acid slightly more thoroughly than the other.
But, like the cosmic rays, the beams of light have a marginal effect. Countless interactions would be needed to leave behind a noticeable imbalance, so some other force may be at work as well. Light would have to pulverize untenably huge quantities of molecules to explain the excesses on its own, Dworkin says.
Sasselov has encouraged Globus and Blandford to consider whether cosmic rays might join forces with polarized light to shape the amino acids on asteroids. On Earth, the doses of cosmic rays — which he likens to supersonic bullets — that would be needed to make a noticeable chiral difference might prove too lethal, he speculated. “You’re destroying so much of everything,” he said. “You may be left with the [correct] handedness, but essentially you’re shooting yourself in the foot.”
Ultimately, the fact that researchers struggle to find a theory that balances the rise of chirality against the destruction of biological materials suggests that our ancestors may have been lucky to find that fine line.
“There is something special about planets like the Earth that protect this kind of chemistry,” Sasselov said.
Imagine this: You have a large multi-module ClojureScript project, and you are planning to make a new deployment in the production. Your project is using CLJS Compiler’s Advanced Optimizations mode. Everything seems to be working smoothly. You are performing some last E2E tests before making a release.
Then, a faulty module loads. BAM! You are slapped in the face with an error something like this:
Uncaught Error: No protocol method IMultiFn.-add-method defined for type function: function XR() { [native code] }
at Nb ((index):964)
at yh ((index):443)
at (index):236
(index):143 Uncaught TypeError: Cannot read property 'j' of undefined
at NU ((index):143)
at login-a5a67c0fffdfa158b7220c8c2553253b645e4e2e.js:169
You did not expect this. There were no significant new changes since the last build, and the unit tests were all hunky-dory.
Additionally, everything worked fine when using Firefox, but now the compiled code crashes on Google Chrome. What is going on? The release deadline is looming and you have to fix this ASAP.
Holy Stack Trace, Batman!
When you build your code with the CLJS compiler, it emits JavaScript code that is compatible with the advanced_optimizations level of the Google Closure compiler. Then, if ":optimizations :advanced" compiler option is enabled, the CLJS compiler will minify the result with Closure Compiler’s Advanced Optimizations.
Closure Compiler’s Advanced Optimizations mode can cause some errors in your build that can be uncomfortable to track down if you are new to Closure Compiler. On top of that, help might be hard to find if you are trying to search the web with the error message because the function names are all mangled up.
It is better to avoid these errors by keeping the Advanced Compilation mode in mind from the start of the project. I’ll describe some common methods to help you to avoid and resolve advanced compilation issues, and show how to resolve the more uncommon issue that was described in the beginning and what is the reason behind it.
Avoiding issues with Advanced Compilation
It is recommended to use ":optimizations :advanced" CLJS compiler option for your production builds. You will greatly reduce your final build size by using more aggressive advanced transformations such as dead code removal and aggressive renaming.
Read more about the advanced transformations at here: Closure Compiler compilation levels
Different build tools such as lein-cljsbuild and shadow-cljs might pass different default options to Closure Compiler. I won’t cover the differences in this blog post, but instead, I will provide some general guidelines on how to avoid issues with advanced compilation in your project.
Keep your externs in a good shape
Extern is a mechanism for declaring names that should not be munged (renamed) by Closure Compiler. When you are using external libraries, make sure that they come bundled with externs if you are using a build tool like lein-cljsbuild. Or, provide externs files of your own if needed. Otherwise, Closure Compiler will munge references to externally defined symbols unintentionally during advanced compilation causing errors found only later when running the compiled code.
To add externs of your own, use a compiler option for example: ":externs ["externs.js"]" and provide an externs.js file in your working directory.
Also, It is good to use ":infer-externs true" compiler option. This option will enable generating externs automatically for JavaScript interop calls.
It is worth to know, that Closure Compiler includes externs for stable JavaScript APIs, but newer features that have an experimental status might not be included yet. The experimental APIs change very quickly, so it makes no sense to include them in the Closure Compiler tool. So, if you are planning to use some experimental features, make sure you add them to your own externs file.
Read more about using external JavaScript libraries from here
Splitting code might cause name collisions
If you are splitting your CLJS code into :modules, you might want to use :rename-prefix "..." compiler option.
Split modules are running in the global JavaScript scope, so they might interfere with other code loaded on the same page (e.g. Google Analytics) and cause unpredictable errors if name collisions occur. When using ":rename-prefix", it is best to use a very short string as a prefix, for example: :rename-prefix "r_"
This will increase the final (gzipped) build size slightly because now each munged global variable will be prefixed with “r_”.
Wrap CLJS output to prevent global scope pollution
Closure Compiler will generate a lot of global variables during advanced compilation that can cause name collisions with other code running on the same page. If you are loading other code on the same page, and you are not splitting code into multiple modules you can utilize the ":output-wrapper true" compiler option. The compiler will wrap the outputted JavaScript code with the default "(function(){…};)()". This will prevent polluting the global JavaScript scope and thus will prevent conflicts with other external code.
However, if you have multiple modules in your project you must add a compiler option :rename-prefix-namespace "...". This enables each module to access the variables defined in the other modules it depends on. This works similarly to the ":rename-prefix" option. The difference is that every global variable will be now scoped under one global variable instead of many. For example, if you use a prefix-namespace like this: :rename-prefix-namespace "P", the compiled code will refer to variable "foo" like "p.foo". This option will also increase the final build size like the ":rename-prefix" option, so think before using it.
There are other interesting compiler options not directly related to advanced optimizations, such as ":fn-invoke-direct" which can be useful for optimizing performance-critical code. You can read more about them in CLJS documentation.
Debugging Advanced Compilation bugs
If you suspect you might have a problem with advanced compilation, often hinted by errors in the browser console that might make no sense, it is best to approach the problem in the following way.
Add two additional compiler options ":pseudo-names true" and ":pretty-print true" for your Advanced Compilation build. Your error will now show a readable name that corresponds to the name in the source code. This will help you to deduce if an extern definition is missing.
Additionally, if your error goes away after enabling the above-mentioned compiler options, it gives you a hint of a possible name collision problem. When there is a problem with a munged variable name colliding with an external variable name, the problem disappears when the munged variable name changes, like what happens when enabling the :pseudo-names option.
Often, if not fixed properly, the name collision issues can arise and then be magically “fixed” when one adds new changes, which in turn causes changes in munged variable names in the build output. It is best to eliminate these problems for good.
Sometimes, something funky can happen
In some rare cases, there might be some advanced compilation issues that are not so easy to avoid. Let’s go back to the beginning. We had an error when running our built web app in Chrome. Firefox had no issues.
Uncaught Error: No protocol method IMultiFn.-add-method defined
for type function: function XR() { [native code] }
at Nb ((index):964)
at yh ((index):443)
at (index):236
The Stack Trace was pointing to a certain multimethod in the CLJS source code. At the first glance, the error might make no sense. The compiled code is trying to invoke -add-method IMultiFn protocol method for a function named XR. When searching for the munged XR in the build output, everything seems to be fine. XR looks good and should work, right? The key thing to notice in the error message is the "[native code]" part of "function XR() { [native code] }". This tells us, that the compiled code is trying to invoke a native browser function XR instead of our munged XR. By chance, the Closure Compiler named our multimethod to XR which happens to collide with the browser provided XR function.
Occasionally you can encounter errors like this when using an older version of Closure Compiler. Developers add new features to web browsers all the time. When a user upgrades his/her web browser there is a chance a new reserved word was added that collides with a munged variable.
It turns out, XR is a reserved word added by the WebXR platform API. When the error occurred “XR” was added in the newest version of Google Chrome, but not yet in Firefox. Newer Closure Compiler versions take this into account by providing an extern for it. It is not always easy to upgrade the ClojureScript version and thus the Closure Compiler used in a project.
In that case, you can fix the problem fast by adding a custom extern of your own:
// externs.jsvarXR={};
Conclusion
Many times, Advanced Compilation issues with CLJS are easy to avoid with proper preparation. On rare occasions, some stranger errors can happen. Thus, test your builds with the current browser versions and learn to debug advanced compilation issues swiftly to prevent wasting precious time.
Microsoft has published on Tuesday two out-of-band security updates to patch two vulnerabilities in the Microsoft Windows Codecs Library.
Tracked as CVE-2020-1425 & CVE-2020-1457, the two bugs only impact Windows 10 and Windows Server 2019 distributions.
In security advisories published today, Microsoft said the two security flaws can be exploited with the help of a specially crafted image file.
If the malformed images are opened inside apps that utilize the built-in Windows Codecs Library to handle multimedia content, then attackers would be allowed to run malicious code on a Windows computer and potentially take over the device.
The two bugs -- described as two remote code execution (RCE) vulnerabilities -- received patches earlier today.
The patches have been deployed to customer systems via an update to the Windows Codecs Library, delivered through the Windows Store app -- not the Windows Update mechanism.
"Customers do not need to take any action to receive the update," Microsoft said.
Redmond said the bugs were privately reported and they haven't been used in the wild before today's patches.
The OS maker said it learned of the bugs after a report from Trend Micro's Zero Day Initiative, a program that intermediates communications between security researchers and larger companies. Microsoft credited Abdul-Aziz Hariri for first discovering these bugs, before passing them to the ZDI team.
from Latest Topic for ZDNet in... https://ift.tt/38iahQX
Alzheimer’s disease (AD) is a devastating neurodegenerative disease characterized by a deterioration in memory and other cognitive domains and the cerebral accumulation of amyloid-Ī² peptide. Advancing age is the most significant risk factor for the disease, with incidence doubling every 5 years after the age of 65.1 Approximately 46.8 million people worldwide currently have AD or a related dementia.2 However, this number is expected to increase exponentially in the coming years, largely as a result of the ageing population.2 The associated economic burden of AD is substantial; in 2015, the total estimated worldwide cost of dementia was US$818 billion, which is projected to rise to US$2.0 trillion by 2030 because of the increased prevalence of AD.2 However, despite decades of research, the etiology of AD remains unknown and there are currently no preventative or disease-modifying treatments.
A growing body of experimental and clinical data implicates the gut microbiome in the pathogenesis of several neurological conditions, including autism spectrum disorder (ASD),3 Parkinson’s disease (PD),4 and multiple sclerosis (MS).5,6 More recently, alterations in gut microbiome composition have been observed in patients with AD,7 suggesting a potential role for the microbiome in AD pathogenesis. This hypothesis has been supported by animal models;8,9 for example, germ-free amyloid-Ī² precursor protein transgenic mice have markedly less cerebral amyloid-Ī² pathology compared with control mice with intestinal bacteria.8 Certain bacteria within the gut microbiota are also capable of secreting large amounts of amyloids and lipopolysaccharides,10 a hallmark feature of AD.
Fecal microbiota transplantation (FMT) is the infusion of fecal material from a healthy donor into the gastrointestinal tract of an individual with disease. This procedure represents the most powerful means of modulating the gut microbiome. The therapy has risen to prominence in the last decade following several outbreaks of severe Clostridioides difficile infection (CDI) in North America and Europe, which were caused in part by the emergence of the hypervirulent NAP1/B1/027 strain. In these difficult-to-treat cases as well as in subsequent randomized, controlled trials, FMT has consistently achieved cure rates >90%.11 Serendipitous improvements following FMT have since been reported in a number of extra-intestinal conditions, including ASD,3 MS,12,13 and myoclonus dystonia.14 To our knowledge, ours is the first report of a patient who experienced rapid improvement in their AD symptoms following FMT for recurrent CDI.
Case report
An 82-year-old man presented for opinion and management of recurrent CDI following hospitalization for methicillin-resistant Staphylococcus aureus pneumonia. The patient had previously failed several courses of antibiotics for CDI, including vancomycin, vancomycin with metronidazole, fidaxomicin, and bezlotoxumab, with relapse confirmed via symptom recurrence and positive stool test.
At the time of presentation, the patient was under the care of his primary care physician and his neurologist for the treatment of AD, following a gradual 5-year decline in memory and cognition. The patient was taking memantine (28 mg once daily) and donepezil (23 mg once daily). The patient’s dementia symptoms included confusion, memory loss, depression, and flattened affect. On his most recent Mini-Mental State Examination (MMSE) administered by the neurologist, the patient scored 20, indicating mild cognitive impairment. This result reflected the gastroenterologist’s findings and was within the expected range for patients with AD. The patient’s wife reported that he no longer appeared to enjoy socializing, and required considerable assistance with basic tasks such as food preparation, bathing, and taking his medication. Neuropsychiatric testing revealed significant impairments in the areas of memory and semantic language abilities, nonverbal learning, and divided attention and response inhibition. His scores in these areas were within the first through fifth percentiles for function.
Following a detailed discussion regarding the potential risks and benefits associated with the procedure, the patient underwent a single 300 mL FMT infusion (per the Borody method)15 using stool from the patient’s 85-year-old wife as a donor. The patient’s wife was intellectually acute, with normal affect and stable mood. Following the procedure, the patient’s CDI symptoms resolved, and repeat stool testing 2 months later was negative.
At the follow-up visit 2 months post-FMT, the patient’s wife reported improvements in the patient’s mental acuity and affect. The MMSE was re-administered by the gastroenterologist (and subsequently by the neurologist) and the patient scored 26, indicating normal cognition. Four months post-FMT, the patient reported continued improvement in memory, with no progression in symptoms. The patient now remembered his daughter’s birthday, which he had not been able to recall previously, and was able to correct the physician’s recollections of his symptoms. Six months post-FMT, the patient reported a marked improvement in mood, was more interactive, and showed more expressive affect. Readministration of the MMSE revealed that the patient’s score had further increased to 29.
Because no study was conducted, no ethics approval or consent was required. Verbal consent was obtained from the patient for the publication of the report; however, this was merely a courtesy because all data have been de-identified.
Discussion
To our knowledge, this is the first report of a case of rapid reversal of AD symptoms in a patient following FMT for recurrent CDI. Improvements in AD symptoms occurred as early as 2 months post-FMT and continued to the 6-month follow-up visit (the date of the last follow-up), with no noted reversion of symptoms. The resolution of symptoms occurred in a stepwise manner: first, there was increased mental acuity and improved affect observed at 2 months post-FMT, and this was followed by marked improvements in memory and mood by 4 and 6 months post-FMT, which were accompanied by more expressive affect. These findings paralleled improvements in the patient’s MMSE scores, which increased to within the range of normal cognition by 2 months post-FMT. Eradication of CDI was also confirmed at the 2-month follow-up visit, via the resolution of symptoms and a negative stool test.
The central role of the gut microbiome in neurological dysfunction has been increasingly recognized, and alterations in gut microbiome composition have consistently been reported in ASD,3 PD,4 and MS.5,6 Rapid and dramatic improvements have also been observed in several of these conditions following manipulation of the gut microbiome, including with FMT, which further supports a causal association. In an open-label study in 18 children with ASD, 2 weeks of vancomycin treatment followed by daily FMT infusions for 7 to 8 weeks resulted in significant improvements in core ASD symptoms, which persisted at the 2-year follow-up.3 Furthermore, in a case series of three patients with MS and underlying gastrointestinal symptoms, 5 to 10 days of FMT infusions resulted in a profound reversal of major neurological symptoms, and led to longstanding disease remission.12 Similarly, a 61-year old female with secondary progressive MS achieved disease stability for over 10 years following FMT for recurrent CDI.13 In addition, in a patient with myoclonus dystonia, significant improvement in tremor and functional deficits were reported following the manipulation of gut microbiota with vancomycin, rifaximin, and metronidazole for chronic diarrhoea.14 In another study, a 71-year old male with longstanding constipation and PD experienced a rapid and dramatic reduction in PD symptoms, including an absence of persistent tremors, glabellar tap reflex, and cogwheel rigidity, when his constipation was treated with vancomycin, metronidazole, and colchicine.16
Gut bacteria may contribute to the pathogenesis of AD via a number of mechanisms. Microbe-mediated bidirectional communication pathways exist between the gut and the central nervous system, and many of the neurotransmitters that regulate mood and cognition in the brain, such as gamma-aminobutyric acid (GABA) and dopamine, are also synthesized and catabolized by gut bacteria. GABA dysfunction has been shown to play a role in AD,17 and abnormally high GABA concentrations have recently been identified in reactive astrocytes of post-mortem AD patients.18 Molecular mimicry is a well-documented strategy that is commonly employed by pathogens to gain a competitive advantage over their host. It occurs when similarities between microbial proteins and host peptides result in the cross-activation of autoreactive T or B cells and a loss of self-tolerance.19 Rheumatic fever after Streptoccocus pyogenes infection is a classic example of this; antigen cross-reactivity between streptococcal M protein and cardiac myosin results in the targeting of myocardial tissue.19 However, a number of other well-documented molecular mimicry events exist, and a recent database study catalogued 261 validated host–microbial mimicry interactions.20 Although the focus to date has been on the role of molecular mimicry as a cause of autoimmune diseases, this mechanism may also be exploited in the development of AD. The progressive accumulation of Ī²-amyloid protein plaques between neurons represents a hallmark of AD. However, the initiating factor responsible for the development and propagation of these prion-like proteins is unknown. Amyloids are secreted by a wide range of pathogenic and non-pathogenic bacteria, and play an important role in cell adhesion and biofilm formation.10 Similarities between the tertiary structure of microbe-derived amyloids and Ī²-amyloid in AD may trigger antigen cross-reactivity, thus potentially initiating the formation of AD. Interestingly, both microbe-derived amyloids (such as curli) and AD-related Ī²-amyloid are recognized by the same Toll-like receptor (TLR)2/TLR1 complex.21 Finally, dysbiosis itself may contribute to the development of AD. A dysbiosis-induced increase in gut permeability may potentially lead to persistent systemic inflammation, disruption of the blood–brain barrier, and ultimately neurodegeneration.
Neuroinflammation plays a central role in the pathogenesis of AD,22,23 and is characterized by elevated levels of pro-inflammatory cytokines,23 nuclear factor ĪŗĪ² signalling,24 and aggregation of activated microglia in damaged areas.25 The influence of the gut microbiome on neuroinflammation in AD remains poorly understood, as do the exact mechanisms of action of FMT in this disease. However, the potent immunomodulatory effects of FMT are well recognized. FMT has been demonstrated to ameliorate chronic intestinal colitis in both humans and animal models via the downregulation of pro-inflammatory cytokines, promotion of anti-inflammatory responses, and inhibition of nuclear factor ĪŗĪ² activity.26–28 In addition, similarly beneficial effects have since been observed in several other conditions with an inflammatory component.13,29,30 Thus, FMT may have a positive effect on cognitive function in AD via alterations in the levels of circulating cytokines. Alternatively, by restoring the previously impaired intestinal barrier function, FMT may prevent the translocation of neuroactive compounds and metabolites within the central nervous system that regulate mood and cognition and contribute to inflammation. However, further studies are urgently required to elucidate the exact mechanisms by which FMT may ameliorate symptoms in AD.
To our knowledge, this is the first documented case of a rapid reversal of AD symptoms in a patient following FMT for recurrent CDI. Although this is only a preliminary report, the remarkable resolution of AD symptoms following FMT for CDI is instructive and adds to the evidence that suggests a causal association between the gut microbiome and neurological dysfunction. Given the probable role of the gut microbiome in the pathogenesis of AD, modulation of the microbiome represents a promising avenue of treatment. A randomized, double-blind, placebo-controlled trial is currently underway to evaluate the efficacy of oral FMT in AD, the results of which are eagerly anticipated.
Acknowledgements
Third-party writing assistance was provided by Jordana Campbell, BSc, of Biocite Medical Writing Services.
Declaration of conflicting interest
Sabine Hazan has a pecuniary interest in Ventura Clinical Trials and ProgenaBiome.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
References
1.
Hirtz, D, Thurman, DJ, Gwinn-Hardy, K, et al. How common are the “common” neurologic disorders? Neurology 2007; 68: 326–337. DOI: 10.1212/01.wnl.0000252807.38124.a3. Google Scholar| Crossref | Medline | ISI
Kang, DW, Adams, JB, Coleman, DM, et al. Long-term benefit of microbiota transfer therapy on autism symptoms and gut microbiota. Sci Rep 2019; 9: 5821. DOI: 10.1038/s41598-019-42183-0. Google Scholar| Crossref | Medline
4.
Minato, T, Maeda, T, Fujisawa, Y, et al. Progression of Parkinson’s disease is associated with gut dysbiosis: two-year follow-up study. PLoS One 2017; 12: e0187307. DOI: 10.1371/journal.pone.0187307. Google Scholar| Crossref | Medline
5.
Umeton, R, Eleftheriou, E, Nedelcu, S, et al. The gut microbiome in relapsing multiple sclerosis patients compared to controls. (P2.355). Neurology 2018; 90: P2.355. Google Scholar
6.
Miyake, S, Kim, S, Suda, W, et al. Dysbiosis in the gut microbiota of patients with multiple sclerosis, with a striking depletion of species belonging to clostridia XIVa and IV clusters. PLoS One 2015; 10: e0137429. DOI: 10.1371/journal.pone.0137429. Google Scholar| Crossref | Medline | ISI
7.
Vogt, NM, Kerby, RL, Dill-McFarland, KA, et al. Gut microbiome alterations in Alzheimer’s disease. Sci Rep 2017; 7: 13537. DOI: 10.1038/s41598-017-13601-y. Google Scholar| Crossref | Medline
8.
Harach, T, Marungruang, N, Duthilleul, N, et al. Reduction of Abeta amyloid pathology in APPPS1 transgenic mice in the absence of gut microbiota. Sci Rep 2017; 7: 41802. DOI: 10.1038/srep41802. Google Scholar| Crossref | Medline | ISI
9.
Minter, MR, Zhang, C, Leone, V, et al. Antibiotic-induced perturbations in gut microbial diversity influences neuro-inflammation and amyloidosis in a murine model of Alzheimer’s disease. Sci Rep 2016; 6: 30028–30028. DOI: 10.1038/srep30028. Google Scholar| Crossref | Medline
10.
Zhou, Y, Blanco, LP, Smith, DR, et al. Bacterial amyloids. Methods Mol Biol 2012; 849: 303–320. DOI: 10.1007/978-1-61779-551-0_21. Google Scholar| Crossref | Medline
11.
Van Nood, E, Vrieze, A, Nieuwdorp, M, et al. Duodenal infusion of donor feces for recurrent clostridium difficile. N Engl J Med 2013; 368: 407–415. DOI: 10.1056/NEJMoa1205037. Google Scholar| Crossref | Medline | ISI
12.
Borody, T, Leis, S, Campbell, J, et al. Fecal microbiota transplantation (FMT) in multiple sclerosis (MS): 942. Am J Gastroenterol 2011; 106: pS352. Google Scholar
13.
Makkawi, S, Camara-Lemarroy, C, Metz, L. Fecal microbiota transplantation associated with 10 years of stability in a patient with SPMS. Neurol Neuroimmunol Neuroinflamm 2018; 5: e459. DOI: 10.1212/NXI.0000000000000459. Google Scholar| Crossref | Medline
14.
Borody, T, Rosen, D, Torres, M, et al. Myoclonus-dystonia affected by GI microbiota? Am J Gastroenterol 2011; 106: S352. Google Scholar
15.
Leis, S, Borody, TJ, Jiang, C, et al. Fecal microbiota transplantation: a ‘How-To’ guide for nurses. Collegian 2015; 22: 445–451. DOI: 10.1016/j.colegn.2014.08.002. Google Scholar| Crossref | Medline
16.
Borody, T, Torres, M, Campbell, J, et al. Treatment of severe constipation improves Parkinson’s disease (PD) symptoms. Am J Gastroenterol 2009; 104: S999. Google Scholar| Crossref
17.
Assefa, BT, Gebre, AK, Altaye, BM. Reactive astrocytes as drug target in Alzheimer’s disease. Biomed Res Int 2018; 2018: 4160247–4160247. DOI: 10.1155/2018/4160247. Google Scholar| Crossref | Medline
Cunningham, MW. Rheumatic fever, autoimmunity, and molecular mimicry: the streptococcal connection. Int Rev Immunol 2014; 33: 314–329. DOI: 10.3109/08830185.2014.917411. Google Scholar| Crossref | Medline
20.
Garg, A, Kumari, B, Kumar, R, et al. miPepBase: a database of experimentally verified peptides involved in molecular mimicry. Front Microbiol 2017; 8: 2053. DOI: 10.3389/fmicb.2017.02053. Google Scholar| Crossref | Medline
21.
Rapsinski, GJ, Newman, TN, Oppong, GO, et al. CD14 protein acts as an adaptor molecule for the immune recognition of Salmonella curli fibers. J Biol Chem 2013; 288: 14178–14188. DOI: 10.1074/jbc.M112.447060. Google Scholar| Crossref | Medline
22.
Azizi, G, Mirshafiey, A. The potential role of proinflammatory and antiinflammatory cytokines in Alzheimer disease pathogenesis. Immunopharmacol Immunotoxicol 2012; 34: 881–895. DOI: 10.3109/08923973.2012.705292. Google Scholar| Crossref | Medline
23.
Su, F, Bai, F, Zhang, Z. Inflammatory cytokines and Alzheimer’s disease: a review from the perspective of genetic polymorphisms. Neurosci Bull 2016; 32: 469–480. DOI: 10.1007/s12264-016-0055-4. Google Scholar| Crossref | Medline
24.
Chen, CH, Zhou, W, Liu, S, et al. Increased NF-ĪŗB signalling up-regulates BACE1 expression and its therapeutic potential in Alzheimer’s disease. Int J Neuropsychopharmacol 2012; 15: 77–90. DOI: 10.1017/S1461145711000149. Google Scholar| Crossref | Medline
Wang, Z, Hua, W, Li, C, et al. Protective role of fecal microbiota transplantation on colitis and colitis-associated colon cancer in mice is associated with Treg cells. Front Microbiol 2019; 10: 2498–2498. DOI: 10.3389/fmicb.2019.02498. Google Scholar| Crossref | Medline
27.
Burrello, C, Garavaglia, F, CribiĆ¹, FM, et al. Therapeutic faecal microbiota transplantation controls intestinal inflammation through IL10 secretion by immune cells. Nat Commun 2018; 9: 5184. DOI: 10.1038/s41467-018-07359-8. Google Scholar| Crossref | Medline
28.
Khoruts, A, Sadowsky, MJ. Understanding the mechanisms of faecal microbiota transplantation. Nat Rev Gastroenterol Hepatol 2016; 13: 508–516. DOI: 10.1038/nrgastro.2016.98. Google Scholar| Crossref | Medline
29.
Chen, D, Wu, J, Jin, D, et al. Fecal microbiota transplantation in cancer management: current status and perspectives. Int J Cancer 2019; 145: 2021–2031. DOI: 10.1002/ijc.32003. Google Scholar| Crossref | Medline
30.
Wang, WW, Zhang, Y, Huang, XB, et al. Fecal microbiota transplantation prevents hepatic encephalopathy in rats with carbon tetrachloride-induced acute hepatic dysfunction. World J Gastroenterol 2017; 23: 6983–6994. DOI: 10.3748/wjg.v23.i38.6983. Google Scholar| Crossref | Medline
Easter Island’s colossal statues loom large—both literally and figuratively—in the popular imagination. The massive heads and torsos dot the landscape like stone sentinels, standing guard over the isle’s treeless, grassy expanse.
The statues have inspired widespread speculation, awe, and wonder for centuries. But the island, called Rapa Nui by its Indigenous people, has also captured the world’s imagination for an entirely different reason.
Rapa Nui is often seen as a cautionary example of societal collapse. In this story, made popular by geographer Jared Diamond’s bestselling book Collapse, the Indigenous people of the island, the Rapanui, so destroyed their environment that, by around 1600, their society fell into a downward spiral of warfare, cannibalism, and population decline. These catastrophes, the collapse narrative explains, resulted in the destruction of the social and political structures that were in place during precolonial times, though the people of Rapa Nui survive and persist on the island to the present day.
In recent years, researchers working on the island have questioned this long-accepted story. For example, anthropologist Terry Hunt and archaeologist Carl Lipo, who have studied the island’s archaeology and cultural history for many years, have suggested an alternative hypothesis that the Rapanui did not succumb to a downward spiral of self-destruction but instead practiced resiliency, cooperation, and perhaps even a degree of environmental stewardship.
Now new evidence from Hunt, Lipo, and their colleagues, published in the Journal of Archaeological Science, lends credence to their ideas. This evidence suggests that the people of the island continued to thrive, as indicated by the continued construction of the stone platforms, called ahu, on which the iconic statues stand, even after the 1600s.
“Our research shows that statue platform construction and use did not end prior to European arrival in 1722,” says Robert DiNapoli, a doctoral student in anthropology at the University of Oregon, who led the study.
This finding, drawing on new statistical methods and excavation work, suggests that the Rapanui were not destitute when the first Europeans arrived. It’s therefore possible that it was the newcomers from Europe who contributed to the island’s societal collapse in the years to come.
Easter Island—known as Rapa Nui by its Indigenous people—features many human-like statues distributed across the isle as shown above. Eric Gaba and Bamse/Wikimedia Commons
The new work is controversial, and not everyone is convinced. But if DiNapoli and his team are correct, the popular story of Rapa Nui’s decline, as described by Diamond, needs to be rethought and its heroes and villains reconfigured. Instead of the Rapanui hastening their own destruction prior to European contact, it is possible that the people of the island may have been the victims of European exploration and exploitation.
Rapa Nui is one of the most remote islands in the world. A tiny speck in the eastern Pacific, it sits more than 2,000 miles west of South America and is about 1,200 miles from its nearest island neighbor, Pitcairn Island.
Archaeologists have documented at least 360 ahu, most of which cluster along the island’s shoreline. They vary in configuration, though most are typically rectangular in shape and are made of basaltic stones neatly fitted together. In addition to their use as statue platforms, the ahu functioned as shrines and places of burial.
DiNapoli and his colleagues used existing radiocarbon dates from previous excavations at 11 different ahu sites. They employed what is called Bayesian analyses, which allow scientists to model the probability of specific events, to build a more precise timeline of construction activities at each site.
The new research indicates that ahu construction began soon after the first Polynesian settlers arrived on the island and continued even after European contact in 1722. This timeline argues against the hypothesized societal collapse occurring around 1600.
The downturn of the islanders, DiNapoli and his colleagues claim, began only after Europeans ushered in a period characterized by disease, murder, slave raiding, and other conflicts.
Not all Rapa Nui specialists agree with DiNapoli and his colleagues’ methods or conclusions. Jo Anne Van Tilburg, an archaeologist at the University of California, Los Angeles, is skeptical that all the radiocarbon dates used by the team reflect specifically ahu-related building events.
Van Tilburg also argues that Diamond’s environmental destruction argument remains a viable hypothesis. “The collapse narrative as these authors describe it is a straw man they have set up that does not accurately reflect the actual hypothesis,” she says.
If Europeans were to blame for the decline of Rapanui society, it would be similar to what happened to Indigenous peoples elsewhere.
In short, Van Tilburg believes the new work is missing some of the nuances of Diamond’s original theory. Diamond never described the collapse as a one-time event, Van Tilburg explains, but rather as a series of events that ultimately resulted in destructive societal changes that were hastened by European contact.
Diamond’s hypothesis is based on a mix of oral tradition, evidence of island deforestation, and the work of previous researchers, such as the Norwegian explorer and ethnologist Thor Heyerdahl. (Heyerdahl gained fame in 1947 for sailing a balsa raft, the Kon-Tiki, to test the theory that South Americans may have colonized Polynesia.)
Early 20th-century oral historians working on Rapa Nui theorized that an internecine clash had occurred between islanders. Heyerdahl later popularized his belief that this warfare, combined with deforestation, resulted in the collapse of the island’s social hierarchies and many traditions, such as the building of stone platforms and statues.
The fate of Rapa Nui has been heatedly debated over the last several years with the development of new theories and innovative techniques, such as Bayesian methods. For many archaeologists, the pre-contact collapse theory is ripe for questioning.
Speaking of the research conducted by DiNapoli’s team, for example, Seth Quintus, an anthropologist at the University of HawaiŹ»i, MÄnoa, who was not involved in the study, says, “Their work adds to the growing body of evidence that has accumulated over the last 10 years that the previous narratives of collapse on Easter Island are not correct—and need to be rethought.”
One mystery the island presents is a lack of trees—an absence that contributes to the belief that ecological devastation led to societal collapse. Theb2b/Wikimedia Commons
If Europeans were to blame for the decline of Rapanui society, that explanation is similar to what happened to other Indigenous peoples elsewhere throughout the world, DiNapoli notes. From that perspective, he says, the popular story of environmental destruction has obscured the islanders’ successes.
“The degree to which their cultural heritage was passed on—and is still present today through language, arts, and cultural practices—is quite notable and impressive,” DiNapoli says. “This degree of resilience has been overlooked due to the collapse narrative and deserves recognition.”
Apple is encouraging developers to prioritize newer web technologies inside their iOS and macOS applications.
In a short technical presentation at the WWDC 2020 conference last week, Apple shared some internal statistics in the hope of convincing app developers to adopt new web technologies and protocols, such as IPv6, HTTP/2, TLS 1.3, and Multipatch TCP.
IPv6
One of the newer technologies that Apple would like developers to implement is IPv6, the next iteration of the IP protocol, set to replace the older IPv4 version.
"Apple platforms have had native IPv6 support for a number of years, including support for IPv6-only networks," said Jiten Mehta, Internet Technologies Engineer at Apple.
"There has been a growing trend of IPv6 usage on the internet. If we look at the last month of connections made worldwide by Apple devices, we see that IPv6 now accounts for 26% of all connections made," Mehta added. "20% of the time, the connection could have used IPv6, but the server didn't have it enabled.
"And when IPv6 is in use, the median connection setup is 1.4 times faster than IPv4. This is primarily due to reduced NAT usage and improved routing."
HTTP/2
Another technology that Apple is trying to get app developers to adopt is HTTP/2, the current and latest version of HTTP, the protocol used to load websites on the internet.
Mehta said that in the last month, around 79% of all web pages loaded in Safari were loaded via an HTTP/2 connection, which were usually 1.8 times faster to load than the older HTTP/1.1 connections.
Apple says that writing apps that work on top of IPv6 and HTTP/2 is now easier than ever. All that app developers have to do is to use the latest versions of the company's networking APIs -- such as URLSession and Network.framework -- which will guarantee the apps will work on IPv6 and HTTP/2 by default, with fallbacks to older protocols, when needed.
Furthermore, Apple is also looking ahead. The company will also add experimental support in macOS 11 and iOS 14 -- set to be released this fall -- for HTTP/3, the next major version of the HTTP protocol, currently undergoing standardization at the IETF (Internet Engineering Task Force).
TLS 1.3 support is, however, disabled by default, and is only included on Apple devices for experimental purposes -- with users/app devs being required to flip a switch for it to work.
TLS 1.3
In addition, Mehta also reported good news on the security front, where the newer TLS 1.3 security protocol has been widely deployed after the company added support for it in iOS 13.4 (November 2019).
Mehta said that over the past month, 49% of all HTTPS network connections on modern Apple devices were running TLS 1.3, where HTTPS connections were being established 1.3 times faster than connections handled via the older TLS 1.2.
Multipath TCP
And the final network technology that Apple hopes app developers chose to support is Multipath TCP, an extension of the Transmission Control Protocol (TCP) that allows connections to use multiple network paths to load the same data.
Mehta said Apple uses this technology to allow apps to continue working without glitches even as users switch their internet network.
At Apple, Mehta said the company has had "great success" using this new technology for Apple Music, where the company reported a 13% reduction in stream stalls and a 22% reduction in stream stall duration, as the Apple Music service was able to function without re-starting downloads when users changed networks.
Last week, at the same WWDC 2020 conference, Apple also announced full support for encrypted DNS on iOS and macOS, with plans to support DNS-over-HTTPS and DNS-over-TLS starting macOS 11 and iOS 14.
from Latest Topic for ZDNet in... https://ift.tt/2YPwgM6
Turing Award Winner and Facebook Chief AI Scientist Yann LeCun has announced his exit from popular social networking platform Twitter after getting involved in a long and often acrimonious dispute regarding racial biases in AI.
Unlike most other artificial intelligence researchers, LeCun has often aired his political views on social media platforms, and has previously engaged in public feuds with colleagues such as Gary Marcus. This time however LeCun’s penchant for debate saw him run afoul of what he termed “the linguistic codes of modern social justice.”
It all started on June 20 with a tweet regarding the new Duke University PULSE AI photo recreation model that had depixelated a low-resolution input image of Barack Obama into a photo of a white male. Penn State University Associate Professor Brad Wyble tweeted “This image speaks volumes about the dangers of bias in AI.” LeCun responded, “ML systems are biased when data is biased. This face upsampling system makes everyone look white because the network was pretrained on FlickFaceHQ, which mainly contains white people pics. Train the *exact* same system on a dataset from Senegal, and everyone will look African.“
Research scientist, co-founder of the “Black in AI” group, and technical co-lead of the Ethical Artificial Intelligence Team at Google Timnit Gebru tweeted in response, “I’m sick of this framing. Tired of it. Many people have tried to explain, many scholars. Listen to us. You can’t just reduce harms caused by ML to dataset bias.” She added, “Even amidst of world wide protests people don’t hear our voices and try to learn from us, they assume they’re experts in everything. Let us lead her and you follow. Just listen. And learn from scholars like @ruha9 [Ruha Benjamin, Associate Professor of African American Studies at Princeton University]. We even brought her to your house, your conference.” (This was a reference to ICLR 2020, where LeCun served as president and Benjamin presented the talk 2020 Vision: Reimagining the Default Settings of Technology & Society.)
Known for her work on racial and gender bias in facial recognition systems and other AI algorithms, Gebru has been advocating for fairness and ethics in AI for years. The Gender Shades project that she leads with MIT Media Lab Computer Scientist Joy Buolamwini revealed that commercial facial recognition software was more likely to misclassified and was less accurate with darker-skinned females compared to lighter-skinned men.
Gebru’s CVPR 2020 talk Computer vision in practice: who is benefiting and who is being harmed? again addressed the role of bias in AI, “I think that now a lot of people have understood that we need to have more diverse datasets, but unfortunately I felt like that’s kind of where the understanding has stopped. It’s like ‘let’s diversify our datasets. And that’s kind of ethics and fairness, right?’ But you can’t ignore social and structural problems.“
LeCun replied that his comment was targeting the particular case of the Duke model and dataset. “The consequences of bias are considerably more dire in a deployed product than in an academic paper,” continued LeCun in a lengthy thread of tweets suggesting it’s not ML researchers that need to be more careful selecting data but engineers.
“Again. UNBELIEVABLE. What does it take? If tutorials at your own conference, books and books and talks and talks from experts coming to YOU, to your own house, feeding it to you, Emily and I even cover issues with how the research community approaches data. Nope. Doesn’t matter.” Gebru replied. “This is not even people outside the community, which we say people like him should follow, read, learn from. This is us trying to educate people in our own community. Its a depressing time to be sure. Depressing.”
Others from the AI and activist communities joined the fray, with far too many simply attacking either LeCun or Gebru. On June 25 LeCun offered an olive branch: “I very much admire your work on AI ethics and fairness. I care deeply about about working to make sure biases don’t get amplified by AI and I’m sorry that the way I communicated here became the story.” Gebru replied, “We’re often told things like ‘I’m sorry that’s how it made you feel.’ That doesn’t really own up to the actual thing. I hope you understand *why* *how* you communicated became the story. It became the story because its a pattern of marginalization.”
The week-long back-and-forth between LeCun and Gebru attracted thousands of likes, comments, and retweets, with a number of high-profile AI researchers expressing dissatisfaction with LeCun’s explanations. Google Research scientist David Ha commented, “I respectfully disagree w/Yann here. As long as progress is benchmarked on biased data, such biases will also be reflected in the inductive biases of ML systems Advancing ML with biased benchmarks and asking engineers to simply ‘retrain models with unbiased data’ is not helpful.” Canada CIFAR AI chair Nicolas Le Roux tweeted, “Yann, I know you mean well. I saw many people act like you just did in good faith, and get defensive when people pointed that this was not the proper response, until one day they stopped to listen and reflect and ultimately change their behaviour.”
Amid the heated debate, the Duke PULSE research team updated their paper, adding: “Overall, it seems that sampling from StyleGAN yields white faces much more frequently than faces of people of color.” The researchers referenced an April 2020 paper on demographic bias in artificially generated facial pictures by Salminen et al.: “Results indicate a racial bias among the generated pictures, with close to three-[fourths] (72.6%) of the pictures representing White people. Asian (13.8%) and Black (10.1%) are considerably less frequent, while Indians represent only a minor fraction of the pictures (3.4%).”
The team also added a “model card” to their study. Gebru was part of a team that introduced the model card framework in 2019 to “provide benchmarked evaluation in a variety of conditions, such as across different cultural, demographic, or phenotypic groups (e.g., race, geographic location, sex, Fitzpatrick skin type) and intersectional groups (e.g., age and race, or sex and Fitzpatrick skin type) that are relevant to the intended application domains.”
Slides from Gebru’s CVPR 2020 tutorial Computer vision in practice: who is benefiting and who is being harmed?
The artificial intelligence community has made a number of moves in recent years to encourage diversity and inclusivity, such as the “AI for All” initiative launched by Gebru’s Stanford supervisor Fei-Fei Li, and scheduling the major AI conference ICLR 2020 in Ethiopia (the conference went virtual due COVID-19). This year, NeurIPS, the world’s most prestigious AI conference, required authors to include a statement of the potential broader impact of their submitted papers, “including its ethical aspects and future societal consequences. Authors should take care to discuss both positive and negative outcomes.”
Journalist: Fangyu Cai | Editor: Michael Sarazen
We know you don’t want to miss any story. Subscribe to our popularSynced Global AI Weeklyto get weekly AI updates.
.JPG){kind=link}