AI Art Video

AI Art Video — independent reviews, comparisons, pricing and step-by-step guides on Aizhi.

Halite AI Programming Competition

Halite is an open-source computer programming contest developed by the hedge fund/tech firm Two Sigma in partnership with a team at Cornell Tech. Programmers can see the game environment and learn everything they need to know about the game. Participants are asked to build bots in whichever language they choose to compete on a two-dimensional virtual battle field. == History == Benjamin Spector and Michael Truell created the first Halite competition in 2016, before partnering with Two Sigma later that year. === Halite I === Halite I asked participants to conquer territory on a grid. It launched in November 2016 and ended in February 2017. Halite I attracted about 1,500 players. === Halite II === Halite II was similar to Halite I, but with a space-war theme. It ran from October 2017 until January 2018. The second installment of the competition attracted about 6,000 individual players from more than 100 countries. Among the participants were professors, physicists and NASA engineers, as well as high school and university students. === Halite III === Halite III launched in mid-October 2018. It ran from October 2018 to January 2019, with an ocean themed playing field. Players were asked to collect and manage Halite, an energy resource. By the end of the competition, Halite III included more than 4000 players and 460 organizations. === Halite IV === Halite IV was hosted by Kaggle, and launched in mid-June 2020.
Read more →
TikTokification

TikTokification (also written TikTok-ification) is a term used to describe the widespread adoption of TikTok's short-form, vertical video format and its algorithmic content-delivery model across the broader social media landscape. The phenomenon encompasses the strategic and cultural changes made by competing platforms such as Instagram, YouTube, Facebook, Snapchat, and LinkedIn in response to TikTok's global dominance. Beyond platform design, the term is also used more broadly to describe shifts in media consumption habits, advertising strategies, and, more critically, the potential cognitive and psychological effects associated with constant short-form video consumption. == Background == === Origins of short-form video === The short-form video format predates TikTok. Vine, launched in 2013, popularised six-second looping videos before shutting down in 2017. TikTok itself, known as Douyin in the Chinese market, was created by the Chinese technology company ByteDance in September 2016. Following its international expansion and its 2018 merger with Musical.ly, TikTok grew rapidly. By 2020, the application had surpassed two billion total downloads worldwide, with over 800 million monthly active users. A key driver of TikTok's success was its recommendation algorithm. The platform's "For You Page" (FYP) serves content to users based on behaviour rather than follower count, making it possible for unknown creators to achieve widespread reach organically. Analysts noted that TikTok serves "fast, visually engaging, and authentic videos that feel more like entertainment than advertising," fundamentally reshaping consumer expectations of digital content. TikTok has been described as "the center of the internet for young people," where users go for entertainment, news, trends, and shopping. As of the mid-2020s, TikTok had approximately 1.12 billion monthly active users. == Platform responses == TikTok's success compelled nearly every major social media platform to restructure its product around short-form video. In 2020, Instagram launched Reels and YouTube launched Shorts, both directly in response to TikTok's growth. Platforms like Meta's Instagram Reels and Google's YouTube Shorts subsequently expanded aggressively, launching new features, creator tools, and even considering separate standalone applications to compete. LinkedIn, traditionally a professional networking site, began experimenting with TikTok-style short-form vertical video feeds. Facebook launched a singular unified video feed combining Reels, long videos, and live videos, similar in structure to TikTok's feed. Snapchat redesigned its application to combine Stories and Spotlight into a unified entertainment feed. YouTube extended its Shorts format to allow videos up to three minutes in length, up from the previous limit of sixty seconds, as of October 2024. Despite these adaptations, experts noted that none of TikTok's rivals had matched its algorithmic precision as of mid-2025. == Societal and cultural impact == === Media and journalism === News organisations have also been affected by TikTokification. Short-form video grew rapidly as a format for news content, driven in large part by TikTok's popularity. According to Pew Research Center, 17% of adults in the United States reported regularly getting news from TikTok in 2024, with 63% of teenagers saying they used the platform as a news source. In response, major publishers began creating bespoke short-form content for TikTok's audience, with organisations such as the BBC building dedicated internal TikTok teams. === Advertising and commerce === TikTokification has had significant effects on the advertising industry. US social video advertising spending was projected to surpass linear television advertising spending for the first time in 2025. Global social commerce sales were projected to reach approximately $900 billion in 2025, with platforms like Douyin and TikTok driving a large share of that growth. TikTok itself generated an estimated $23.6 billion in advertising revenue in 2024. Short-form video has been described as bridging the gap between brand awareness and direct conversion. Surveys have found that consumers trust user-generated content 8.7 times more than influencer content and 6.6 times more than branded content, prompting brands to favour creator-led video formats. === Attention spans and cognitive effects === A growing body of research has examined the cognitive consequences of heavy short-form video consumption, a set of effects sometimes referred to as "TikTok Brain." A large systematic review and meta-analysis published in Psychological Bulletin, analysing data from 98,299 participants across 71 studies, found that the more short-form video content a person watches, the poorer their cognitive performance in attention and inhibitory control. The review also found that greater engagement with short-form video was associated with higher levels of anxiety, depression, and stress, as well as sleep disturbances. The platform's inherent demand for engaging content has resulted in the proliferation of sludge content, a genre of split screen video with the main video on the top and an unrelated attention-grabbing video on the bottom, typically repetitive gameplay (notably of the endless runner mobile game Subway Surfers) or oddly satisfying videos, designed to maximize viewer retention in cases where the main video may appear uninteresting and would normally cause the viewer to skip it. Sludge content is often described as overstimulating, reflecting and contributing to declining attention spans, though the scholarly evidence supporting such claims is not conclusive. Dr. Yann Poncin, associate professor at the Child Study Center at Yale University, noted that "infinite scrolling and short-form video are designed to capture your attention in short bursts," contrasting this with earlier entertainment formats that guided audiences through longer narratives. Research suggests that children and teenagers may be particularly vulnerable, with early exposure to rapid frame changes potentially conditioning the brain's neural pathways to require constant stimulation, making it more challenging to engage with slower-paced activities. A separate study published in Nature Communications by researchers at the Technical University of Denmark documented a notable decrease in collective attention span over time, attributing it in part to the increasing volume and pace of content production and consumption online. Researchers caution, however, that the majority of relevant studies are cross-sectional, meaning they capture data at a single point in time and cannot establish causality. It remains possible that individuals with pre-existing conditions such as anxiety or attention deficits may be more likely to engage heavily with these platforms as a coping mechanism. === Academic and sociological analysis === Scholars have framed TikTokification within the context of the attention economy. A 2024 academic analysis described TikTok as representing "a new paradigm of social media communication" shaped by youth culture, mobile technology, and the economics of attention, in which spectators become active contributors to a shared content pipeline. The same analysis noted that TikTok "reflects a new mode of communication influenced by avant-garde cinema, the use of mobile technology, and the social habits of particular social groups." US social media users were projected to spend 61.1% of their time on social networks watching videos in 2025, up from 33.3% in 2019, before TikTok became widely popular, underscoring the scale of the behavioural shift. == Monetisation challenges == Despite high engagement levels, monetising short-form video has remained difficult for platforms and creators alike. Unlike long-form YouTube content, short clips offer limited space for advertisers to insert advertisements. YouTube Shorts pays approximately four cents per 1,000 views, considerably less than its long-form counterpart. From 2025 onward, platforms began introducing creator funds, advertisements, and AI-driven content recommendations as part of broader efforts to make short-form video economically sustainable for creators.
Read more →
Big memory

Big-memory computers are machines with a large amount of random-access memory (RAM). The computers are required for databases, graph analytics, or more generally, high-performance computing, data science, and big data. Some database systems called in-memory databases are designed to run mostly in memory, rarely if ever retrieving data from disk or flash memory. See list of in-memory databases. == Details == The performance of big-memory systems depends on how the central processing units (CPUs) access the memory, via a conventional memory controller or via non-uniform memory access (NUMA). Performance also depends on the size and design of the CPU cache. Performance also depends on operating system (OS) design. The huge pages feature in Linux and other OSes can improve the efficiency of virtual memory. The transparent huge pages feature in Linux can offer better performance for some big-memory workloads. The "Large-Page Support" in Microsoft Windows enables server applications to establish large-page memory regions which are typically three orders of magnitude larger than the native page size.
Read more →
Client-side encryption

Client-side encryption is the cryptographic technique of encrypting data on the sender's side, before it is transmitted to a server such as a cloud storage service. Client-side encryption features an encryption key that is not available to the service provider, making it difficult or impossible for service providers to decrypt hosted data. Client-side encryption allows for the creation of applications whose providers cannot access the data its users have stored, thus offering a high level of privacy. Applications utilizing client-side encryption are sometimes marketed under the misleading or incorrect term "zero-knowledge", but this is a misnomer, as the term zero-knowledge describes something entirely different in the context of cryptography. == Details == Client-side encryption seeks to eliminate the potential for data to be viewed by service providers (or third parties that compel service providers to deliver access to data), client-side encryption ensures that data and files that are stored in the cloud can only be viewed on the client-side of the exchange. This prevents data loss and the unauthorized disclosure of private or personal files, providing increased peace of mind for its users. Current recommendations by industry professionals as well as academic scholars offer great vocal support for developers to include client-side encryption to protect the confidentiality and integrity of information. === Examples of services that use client-side encryption by default === Tresorit MEGA Cryptee Cryptomator === Examples of services that optionally support client-side encryption === Apple iCloud offers optional client-side encryption when "Advanced Data Protection for iCloud" is enabled. Google Drive, Google Docs, Google Meet, Google Calendar, and Gmail — However, as of Jul 2024, optional client-side encryption features are only available to paid users. === Examples of services that do not support client-side encryption === Dropbox === Examples of client-side encrypted services that no longer exist === SpiderOak Backup
Read more →
Gödel machine

A Gödel machine is a hypothetical self-improving computer program that solves problems in an optimal way. It uses a recursive self-improvement protocol in which it rewrites its own code when it can prove the new code provides a better strategy. The machine was invented by Jürgen Schmidhuber (first proposed in 2003), but is named after Kurt Gödel who inspired the mathematical theories. The Gödel machine is often discussed when dealing with issues of meta-learning, also known as "learning to learn." Applications include automating human design decisions and transfer of knowledge between multiple related tasks, and may lead to design of more robust and general learning architectures. Though theoretically possible, no full implementation has been created. The Gödel machine is often compared with Marcus Hutter's AIXI, another formal specification for an artificial general intelligence. Schmidhuber points out that the Gödel machine could start out by implementing AIXItl as its initial sub-program, and self-modify after it finds proof that another algorithm for its search code will be better. == Limitations == Traditional problems solved by a computer only require one input and provide some output. Computers of this sort had their initial algorithm hardwired. This does not take into account the dynamic natural environment, and thus was a goal for the Gödel machine to overcome. The Gödel machine has limitations of its own, however. According to Gödel's First Incompleteness Theorem, any formal system that encompasses arithmetic is either flawed or allows for statements that cannot be proved in the system. Hence even a Gödel machine with unlimited computational resources must ignore those self-improvements whose effectiveness it cannot prove. == Variables of interest == There are three variables that are particularly useful in the run time of the Gödel machine. At some time t {\displaystyle t} , the variable time {\displaystyle {\text{time}}} will have the binary equivalent of t {\displaystyle t} . This is incremented steadily throughout the run time of the machine. Any input meant for the Gödel machine from the natural environment is stored in variable x {\displaystyle x} . It is likely the case that x {\displaystyle x} will hold different values for different values of variable time {\displaystyle {\text{time}}} . The outputs of the Gödel machine are stored in variable y {\displaystyle y} , where y ( t ) {\displaystyle y(t)} would be the output bit-string at some time t {\displaystyle t} . At any given time t {\displaystyle t} , where ( 1 ≤ t ≤ T ) {\displaystyle (1\leq t\leq T)} , the goal is to maximize future success or utility. A typical utility function follows the pattern u ( s , E n v ) : S × E → R {\displaystyle u(s,\mathrm {Env} ):S\times E\rightarrow \mathbb {R} } : u ( s , E n v ) = E μ [ ∑ τ = time T r ( τ ) ∣ s , E n v ] {\displaystyle u(s,\mathrm {Env} )=E_{\mu }{\Bigg [}\sum _{\tau ={\text{time}}}^{T}r(\tau )\mid s,\mathrm {Env} {\Bigg ]}} where r ( t ) {\displaystyle r(t)} is a real-valued reward input (encoded within s ( t ) {\displaystyle s(t)} ) at time t {\displaystyle t} , E μ [ ⋅ ∣ ⋅ ] {\displaystyle E_{\mu }[\cdot \mid \cdot ]} denotes the conditional expectation operator with respect to some possibly unknown distribution μ {\displaystyle \mu } from a set M {\displaystyle M} of possible distributions ( M {\displaystyle M} reflects whatever is known about the possibly probabilistic reactions of the environment), and the above-mentioned time = time ⁡ ( s ) {\displaystyle {\text{time}}=\operatorname {time} (s)} is a function of state s {\displaystyle s} which uniquely identifies the current cycle. Note that we take into account the possibility of extending the expected lifespan through appropriate actions. == Instructions used by proof techniques == The nature of the six proof-modifying instructions below makes it impossible to insert an incorrect theorem into proof, thus trivializing proof verification. === get-axiom(n) === Appends the n-th axiom as a theorem to the current theorem sequence. Below is the initial axiom scheme: Hardware Axioms formally specify how components of the machine could change from one cycle to the next. Reward Axioms define the computational cost of hardware instruction and the physical cost of output actions. Related Axioms also define the lifetime of the Gödel machine as scalar quantities representing all rewards/costs. Environment Axioms restrict the way new inputs x are produced from the environment, based on previous sequences of inputs y. Uncertainty Axioms/String Manipulation Axioms are standard axioms for arithmetic, calculus, probability theory, and string manipulation that allow for the construction of proofs related to future variable values within the Gödel machine. Initial State Axioms contain information about how to reconstruct parts or all of the initial state. Utility Axioms describe the overall goal in the form of utility function u. === apply-rule(k, m, n) === Takes in the index k of an inference rule (such as Modus tollens, Modus ponens), and attempts to apply it to the two previously proved theorems m and n. The resulting theorem is then added to the proof. === delete-theorem(m) === Deletes the theorem stored at index m in the current proof. This helps to mitigate storage constraints caused by redundant and unnecessary theorems. Deleted theorems can no longer be referenced by the above apply-rule function. === set-switchprog(m, n) === Replaces switchprog S pm:n, provided it is a non-empty substring of S p. === check() === Verifies whether the goal of the proof search has been reached. A target theorem states that given the current axiomatized utility function u (Item 1f), the utility of a switch from p to the current switchprog would be higher than the utility of continuing the execution of p (which would keep searching for alternative switchprogs). === state2theorem(m, n) === Takes in two arguments, m and n, and attempts to convert the contents of Sm:n into a theorem. == Example applications == === Time-limited NP-hard optimization === The initial input to the Gödel machine is the representation of a connected graph with a large number of nodes linked by edges of various lengths. Within given time T it should find a cyclic path connecting all nodes. The only real-valued reward will occur at time T. It equals 1 divided by the length of the best path found so far (0 if none was found). There are no other inputs. The by-product of maximizing expected reward is to find the shortest path findable within the limited time, given the initial bias. === Fast theorem proving === Prove or disprove as quickly as possible that all even integers > 2 are the sum of two primes (Goldbach’s conjecture). The reward is 1/t, where t is the time required to produce and verify the first such proof. === Maximizing expected reward with bounded resources === A cognitive robot that needs at least 1 liter of gasoline per hour interacts with a partially unknown environment, trying to find hidden, limited gasoline depots to occasionally refuel its tank. It is rewarded in proportion to its lifetime, and dies after at most 100 years or as soon as its tank is empty or it falls off a cliff, and so on. The probabilistic environmental reactions are initially unknown but assumed to be sampled from the axiomatized Speed Prior, according to which hard-to-compute environmental reactions are unlikely. This permits a computable strategy for making near-optimal predictions. One by-product of maximizing expected reward is to maximize expected lifetime.
Read more →
Data independence

Data independence is the type of data transparency that matters for a centralized DBMS. It refers to the immunity of user applications to changes made in the definition and organization of data. Application programs should not, ideally, be exposed to details of data representation and storage. The DBMS provides an abstract view of the data that hides such details. There are two types of data independence: physical and logical data independence. The data independence and operation independence together gives the feature of data abstraction. There are two levels of data independence. == Logical data independence == The logical structure of the data is known as the 'schema definition'. In general, if a user application operates on a subset of the attributes of a relation, it should not be affected later when new attributes are added to the same relation. Logical data independence indicates that the conceptual schema can be changed without affecting the existing schemas. == Physical data independence == The physical structure of the data is referred to as "physical data description". Physical data independence deals with hiding the details of the storage structure from user applications. The application should not be involved with these issues since, conceptually, there is no difference in the operations carried out against the data. There are three types of data independence: Logical data independence: The ability to change the logical (conceptual) schema without changing the External schema (User View) is called logical data independence. For example, the addition or removal of new entities, attributes, or relationships to the conceptual schema or having to rewrite existing application programs. Physical data independence: The ability to change the physical schema without changing the logical schema is called physical data independence. For example, a change to the internal schema, such as using different file organization or storage structures, storage devices, or indexing strategy, should be possible without having to change the conceptual or external schemas. View level data independence: always independent no effect, because there doesn't exist any other level above view level. == Data independence == Data independence can be explained as follows: Each higher level of the data architecture is immune to changes of the next lower level of the architecture. The logical scheme stays unchanged even though the storage space or type of some data is changed for reasons of optimization or reorganization. In this, external schema does not change. In this, internal schema changes may be required due to some physical schema were reorganized here. Physical data independence is present in most databases and file environment in which hardware storage of encoding, exact location of data on disk, merging of records, so on this are hidden from user. == Data independence types == The ability to modify schema definition in one level without affecting schema of that definition in the next higher level is called data independence. There are two levels of data independence, they are Physical data independence and Logical data independence. Physical data independence is the ability to modify the physical schema without causing application programs to be rewritten. Modifications at the physical level are occasionally necessary to improve performance. It means we change the physical storage/level without affecting the conceptual or external view of the data. The new changes are absorbed by mapping techniques. Logical data independence is the ability to modify the logical schema without causing application programs to be rewritten. Modifications at the logical level are necessary whenever the logical structure of the database is altered (for example, when money-market accounts are added to banking system). Logical Data independence means if we add some new columns or remove some columns from table then the user view and programs should not change. For example: consider two users A & B. Both are selecting the fields "EmployeeNumber" and "EmployeeName". If user B adds a new column (e.g. salary) to his table, it will not affect the external view for user A, though the internal schema of the database has been changed for both users A & B. Logical data independence is more difficult to achieve than physical data independence, since application programs are heavily dependent on the logical structure of the data that they access.
Read more →
Internet Security Alliance

Internet Security Alliance (ISA) was founded in 2001 as a non-profit collaboration between Carnegie Mellon University's CyLab and Electronic Industries Alliance, a federation of trade associations. The Internet Security Alliance is focused on cyber security, acting as a forum for information sharing and leadership on information security, and lobbying for corporate security interests. == International operations == The Internet Security Alliance operates with a global membership to provide international security for its partners. The organization's membership includes companies located on four continents, and the Executive Committee always includes at least one non-U.S.-based company. The Internet Security Alliance believes that international communication is crucial for long-term greater information security, as it allows for a more realistic approach to addressing the many challenges faced by users of the Internet. == Publications == Published in 2009, The Financial Impact of Cyber Risk is the first known guidance document to attempt to approach the financial impact of cyber risks from the perspective of core business functions. It claims to provide guidance to CFOs and their colleagues responsible for legal issues, business operations and technology, privacy and compliance, risk assessment and insurance, and corporate communications.
Read more →
Time-lock puzzle

A time-lock puzzle, or time-released cryptography, encrypts a message that cannot be decrypted until a specified amount of time has passed. The concept was first described by Timothy C. May, and a solution first introduced by Ron Rivest, Adi Shamir, and David A. Wagner in 1996. Time-lock puzzle are useful in cases where confidentiality of information is determined by time, such as a diarist who does not want their views released until 50 years after their death, an auction where bids are sealed until the bidding period is closed, electronic voting, and contract signing. They can additionally be used in creating further cryptographic primitives, such as verifiable delay functions and zero knowledge proofs. Time-released cryptography can be achieved through several different mechanisms. Use mathematical problems requiring sequential calculations to solve, and cannot be solved with parallelization. Thus, adding more computers to a problem will not help solve the problem faster. Use of a trusted agent, or multiple agents who each hold a part of the message and cryptographic keys, who release the message after a specified time period has passed. Distribute public encryption keys to users, and place private cryptographic keys with a trusted agent in an offline location, to be released at a later date.
Read more →
Continuous Function Chart

A Continuous Function Chart (CFC) is a graphic editor that can be used in conjunction with the STEP 7 software package or with other tools, such as CODESYS. It is used to create the entire software structure of the CPU from ready-made blocks. When working with the editor, you place blocks on function charts, assign parameters to them, and interconnect them. Interconnecting means, for example, that values are transferred from one output to one or more inputs during communication between the blocks. Continuous function charts are basically used for controlling continuous processes, where all the logic is executed and outputs are calculated in each PLC scan. Whereas in SFC, execution will be sequential as done is batch processes.
Read more →
Trace zero cryptography

First proposed by Gerhard Frey in 1998, trace zero cryptography refers to the use of trace zero varieties (TZV) for cryptographic purpose. Trace zero varieties are subgroups of the divisor class group on a low genus hyperelliptic curve defined over a finite field. These groups can be used to establish asymmetric cryptography using the discrete logarithm problem as cryptographic primitive. Trace zero varieties feature a better scalar multiplication performance than elliptic curves. This allows fast arithmetic in these groups, which can speed up the calculations with a factor 3 compared with elliptic curves and hence speed up the cryptosystem. Another advantage is that for groups of cryptographically relevant size, the order of the group can simply be calculated using the characteristic polynomial of the Frobenius endomorphism. This is not the case, for example, in elliptic curve cryptography when the group of points of an elliptic curve over a prime field is used for cryptographic purpose. However, to represent an element of the trace zero variety more bits are needed compared with elements of elliptic or hyperelliptic curves. Another disadvantage is the fact that it is possible to reduce the security of the TZV of 1/6th of the bit length using cover attack. == Mathematical background == A hyperelliptic curve C of genus g over a prime field F q {\displaystyle \mathbb {F} _{q}} where q = pn (p prime) of odd characteristic is defined as C : y 2 + h ( x ) y = f ( x ) , {\displaystyle C:~y^{2}+h(x)y=f(x),} where f monic, deg(f) = 2g + 1 and deg(h) ≤ g. The curve has at least one F q {\displaystyle \mathbb {F} _{q}} -rational Weierstraßpoint. The Jacobian variety J C ( F q n ) {\displaystyle J_{C}(\mathbb {F} _{q^{n}})} of C is for all finite extension F q n {\displaystyle \mathbb {F} _{q^{n}}} isomorphic to the ideal class group Cl ⁡ ( C / F q n ) {\displaystyle \operatorname {Cl} (C/\mathbb {F} _{q^{n}})} . With the Mumford's representation it is possible to represent the elements of J C ( F q n ) {\displaystyle J_{C}(\mathbb {F} _{q^{n}})} with a pair of polynomials [u, v], where u, v ∈ F q n [ x ] {\displaystyle \mathbb {F} _{q^{n}}[x]} . The Frobenius endomorphism σ is used on an element [u, v] of J C ( F q n ) {\displaystyle J_{C}(\mathbb {F} _{q^{n}})} to raise the power of each coefficient of that element to q: σ([u, v]) = [uq(x), vq(x)]. The characteristic polynomial of this endomorphism has the following form: χ ( T ) = T 2 g + a 1 T 2 g − 1 + ⋯ + a g T g + ⋯ + a 1 q g − 1 T + q g , {\displaystyle \chi (T)=T^{2g}+a_{1}T^{2g-1}+\cdots +a_{g}T^{g}+\cdots +a_{1}q^{g-1}T+q^{g},} where ai in Z {\displaystyle \mathbb {Z} } With the Hasse–Weil theorem it is possible to receive the group order of any extension field F q n {\displaystyle \mathbb {F} _{q^{n}}} by using the complex roots τi of χ(T): | J C ( F q n ) | = ∏ i = 1 2 g ( 1 − τ i n ) {\displaystyle |J_{C}(\mathbb {F} _{q^{n}})|=\prod _{i=1}^{2g}(1-\tau _{i}^{n})} Let D be an element of the J C ( F q n ) {\displaystyle J_{C}(\mathbb {F} _{q^{n}})} of C, then it is possible to define an endomorphism of J C ( F q n ) {\displaystyle J_{C}(\mathbb {F} _{q^{n}})} , the so-called trace of D: Tr ⁡ ( D ) = ∑ i = 0 n − 1 σ i ( D ) = D + σ ( D ) + ⋯ + σ n − 1 ( D ) {\displaystyle \operatorname {Tr} (D)=\sum _{i=0}^{n-1}\sigma ^{i}(D)=D+\sigma (D)+\cdots +\sigma ^{n-1}(D)} Based on this endomorphism one can reduce the Jacobian variety to a subgroup G with the property, that every element is of trace zero: G = { D ∈ J C ( F q n ) | Tr ( D ) = 0 } , ( 0 neutral element in J C ( F q n ) {\displaystyle G=\{D\in J_{C}(\mathbb {F} _{q^{n}})~|~{\text{Tr}}(D)={\textbf {0}}\},~~~({\textbf {0}}{\text{ neutral element in }}J_{C}(\mathbb {F} _{q^{n}})} G is the kernel of the trace endomorphism and thus G is a group, the so-called trace zero (sub)variety (TZV) of J C ( F q n ) {\displaystyle J_{C}(\mathbb {F} _{q^{n}})} . The intersection of G and J C ( F q ) {\displaystyle J_{C}(\mathbb {F} _{q})} is produced by the n-torsion elements of J C ( F q ) {\displaystyle J_{C}(\mathbb {F} _{q})} . If the greatest common divisor gcd ( n , | J C ( F q ) | ) = 1 {\displaystyle \gcd(n,|J_{C}(\mathbb {F} _{q})|)=1} the intersection is empty and one can compute the group order of G: | G | = | J C ( F q n ) | | J C ( F q ) | = ∏ i = 1 2 g ( 1 − τ i n ) ∏ i = 1 2 g ( 1 − τ i ) {\displaystyle |G|={\dfrac {|J_{C}(\mathbb {F} _{q^{n}})|}{|J_{C}(\mathbb {F} _{q})|}}={\dfrac {\prod _{i=1}^{2g}(1-\tau _{i}^{n})}{\prod _{i=1}^{2g}(1-\tau _{i})}}} The actual group used in cryptographic applications is a subgroup G0 of G of a large prime order l. This group may be G itself. There exist three different cases of cryptographical relevance for TZV: g = 1, n = 3 g = 1, n = 5 g = 2, n = 3 == Arithmetic == The arithmetic used in the TZV group G0 based on the arithmetic for the whole group J C ( F q n ) {\displaystyle J_{C}(\mathbb {F} _{q^{n}})} , But it is possible to use the Frobenius endomorphism σ to speed up the scalar multiplication. This can be archived if G0 is generated by D of order l then σ(D) = sD, for some integers s. For the given cases of TZV s can be computed as follows, where ai come from the characteristic polynomial of the Frobenius endomorphism : For g = 1, n = 3: s = q − 1 1 − a 1 mod ℓ {\displaystyle s={\dfrac {q-1}{1-a_{1}}}{\bmod {\ell }}} For g = 1, n = 5: s = q 2 − q − a 1 2 q + a 1 q + 1 q − 2 a 1 q + a 1 3 − a 1 2 + a 1 − 1 mod ℓ {\displaystyle s={\dfrac {q^{2}-q-a_{1}^{2}q+a_{1}q+1}{q-2a_{1}q+a_{1}^{3}-a_{1}^{2}+a_{1}-1}}{\bmod {\ell }}} For g = 2, n = 3: s = − q 2 − a 2 + a 1 a 1 q − a 2 + 1 mod ℓ {\displaystyle s=-{\dfrac {q^{2}-a_{2}+a_{1}}{a_{1}q-a_{2}+1}}{\bmod {\ell }}} Knowing this, it is possible to replace any scalar multiplication mD (|m| ≤ l/2) with: m 0 D + m 1 σ ( D ) + ⋯ + m n − 1 σ n − 1 ( D ) , where m i = O ( ℓ 1 / ( n − 1 ) ) = O ( q g ) {\displaystyle m_{0}D+m_{1}\sigma (D)+\cdots +m_{n-1}\sigma ^{n-1}(D),~~~~{\text{where }}m_{i}=O(\ell ^{1/(n-1)})=O(q^{g})} With this trick the multiple scalar product can be reduced to about 1/(n − 1)th of doublings necessary for calculating mD, if the implied constants are small enough. == Security == The security of cryptographic systems based on trace zero subvarieties according to the results of the papers comparable to the security of hyper-elliptic curves of low genus g' over F p ′ {\displaystyle \mathbb {F} _{p'}} , where p' ~ (n − 1)(g/g' ) for |G| ~128 bits. For the cases where n = 3, g = 2 and n = 5, g = 1 it is possible to reduce the security for at most 6 bits, where |G| ~ 2256, because one can not be sure that G is contained in a Jacobian of a curve of genus 6. The security of curves of genus 4 for similar fields are far less secure. == Cover attack on a trace zero crypto-system == The attack published in shows, that the DLP in trace zero groups of genus 2 over finite fields of characteristic diverse than 2 or 3 and a field extension of degree 3 can be transformed into a DLP in a class group of degree 0 with genus of at most 6 over the base field. In this new class group the DLP can be attacked with the index calculus methods. This leads to a reduction of the bit length 1/6th.
Read more →
Backup

In information technology, a backup, or data backup is a copy of computer data taken and stored elsewhere so that it may be used to restore the original after a data loss event. The verb form, referring to the process of doing so, is "back up", whereas the noun and adjective form is "backup". Backups can be used to recover data after its loss from data deletion or corruption, or to recover data from an earlier time. Backups provide a simple form of IT disaster recovery; however not all backup systems are able to reconstitute a computer system or other complex configuration such as a computer cluster, active directory server, or database server. A backup system contains at least one copy of all data considered worth saving. The data storage requirements can be large. An information repository model may be used to provide structure to this storage. There are different types of data storage devices used for copying backups of data that is already in secondary storage onto archive files. There are also different ways these devices can be arranged to provide geographic dispersion, data security, and portability. Data is selected, extracted, and manipulated for storage. The process can include methods for dealing with live data, including open files, as well as compression, encryption, and de-duplication. Additional techniques apply to enterprise client-server backup. Backup schemes may include dry runs that validate the reliability of the data being backed up. There are limitations and human factors involved in any backup scheme. == Storage == A backup strategy requires an information repository, "a secondary storage space for data" that aggregates backups of data "sources". The repository could be as simple as a list of all backup media (DVDs, etc.) and the dates produced, or could include a computerized index, catalog, or relational database. === 3-2-1 Backup Rule === The backup data needs to be stored, requiring a backup rotation scheme, which is a system of backing up data to computer media that limits the number of backups of different dates retained separately, by appropriate re-use of the data storage media by overwriting of backups no longer needed. The scheme determines how and when each piece of removable storage is used for a backup operation and how long it is retained once it has backup data stored on it. The 3-2-1 rule can aid in the backup process. It states that there should be at least 3 copies of the data, stored on 2 different types of storage media, and one copy should be kept offsite, in a remote location (this can include cloud storage). 2 or more different media should be used to eliminate data loss due to similar reasons (for example, optical discs may tolerate being underwater while LTO tapes may not, and SSDs cannot fail due to head crashes or damaged spindle motors since they do not have any moving parts, unlike hard drives). An offsite copy protects against fire, theft of physical media (such as tapes or discs) and natural disasters like floods and earthquakes. Physically protected hard drives are an alternative to an offsite copy, but they have limitations like only being able to resist fire for a limited period of time, so an offsite copy still remains as the ideal choice. Because there is no perfect storage, many backup experts recommend maintaining a second copy on a local physical device, even if the data is also backed up offsite. === Backup methods === ==== Unstructured ==== An unstructured repository may simply be a stack of tapes, DVD-Rs or external HDDs with minimal information about what was backed up and when. This method is the easiest to implement, but unlikely to achieve a high level of recoverability as it lacks automation. ==== Full only/System imaging ==== A repository using this backup method contains complete source data copies taken at one or more specific points in time. Copying system images, this method is frequently used by computer technicians to record known good configurations. However, imaging is generally more useful as a way of deploying a standard configuration to many systems rather than as a tool for making ongoing backups of diverse systems. ==== Incremental ==== An incremental backup stores data changed since a reference point in time. Duplicate copies of unchanged data are not copied. Typically a full backup of all files is made once or at infrequent intervals, serving as the reference point for an incremental repository. Subsequently, a number of incremental backups are made after successive time periods. Restores begin with the last full backup and then apply the incrementals. Some backup systems can create a synthetic full backup from a series of incrementals, thus providing the equivalent of frequently doing a full backup. When done to modify a single archive file, this speeds restores of recent versions of files. ==== Near-CDP ==== Continuous Data Protection (CDP) refers to a backup that instantly saves a copy of every change made to the data. This allows restoration of data to any point in time and is the most comprehensive and advanced data protection. Near-CDP backup applications—often marketed as "CDP"—automatically take incremental backups at a specific interval, for example every 15 minutes, one hour, or 24 hours. They can therefore only allow restores to an interval boundary. Near-CDP backup applications use journaling and are typically based on periodic "snapshots", read-only copies of the data frozen at a particular point in time. Near-CDP (except for Apple Time Machine) intent-logs every change on the host system, often by saving byte or block-level differences rather than file-level differences. This backup method differs from simple disk mirroring in that it enables a roll-back of the log and thus a restoration of old images of data. Intent-logging allows precautions for the consistency of live data, protecting self-consistent files but requiring applications "be quiesced and made ready for backup." Near-CDP is more practicable for ordinary personal backup applications, as opposed to true CDP, which must be run in conjunction with a virtual machine or equivalent and is therefore generally used in enterprise client-server backups. Software may create copies of individual files such as written documents, multimedia projects, or user preferences, to prevent failed write events caused by power outages, operating system crashes, or exhausted disk space, from causing data loss. A common implementation is an appended ".bak" extension to the file name. ==== Reverse incremental ==== A Reverse incremental backup method stores a recent archive file "mirror" of the source data and a series of differences between the "mirror" in its current state and its previous states. A reverse incremental backup method starts with a non-image full backup. After the full backup is performed, the system periodically synchronizes the full backup with the live copy, while storing the data necessary to reconstruct older versions. This can either be done using hard links—as Apple Time Machine does, or using binary diffs. ==== Differential ==== A differential backup saves only the data that has changed since the last full backup. This means a maximum of two backups from the repository are used to restore the data. However, as time from the last full backup (and thus the accumulated changes in data) increases, so does the time to perform the differential backup. Restoring an entire system requires starting from the most recent full backup and then applying just the last differential backup. A differential backup copies files that have been created or changed since the last full backup, regardless of whether any other differential backups have been made since, whereas an incremental backup copies files that have been created or changed since the most recent backup of any type (full or incremental). Changes in files may be detected through a more recent date/time of last modification file attribute, and/or changes in file size. Other variations of incremental backup include multi-level incrementals and block-level incrementals that compare parts of files instead of just entire files. === Storage media === Regardless of the repository model that is used, the data has to be copied onto an archive file data storage medium. The medium used is also referred to as the type of backup destination. ==== Magnetic tape ==== Magnetic tape was for a long time the most commonly used medium for bulk data storage, backup, archiving, and interchange. It was previously a less expensive option, but this is no longer the case for smaller amounts of data. Tape is a sequential access medium, so the rate of continuously writing or reading data can be very fast. While tape media itself has a low cost per space, tape drives are typically dozens of times as expensive as hard disk drives and optical drives. Tape media are generally rotated on a schedule so at least one set is off-site in case something should happe
Read more →
Social media coverage of the Olympics

Over the years, television broadcast rights have distinguished what Olympic-related content can be accessed by fans online. By doing so, mobile-friendly social platforms began to integrate into the Olympics. Athletes and fans use these platforms to share live updates, special moments, and behind-the-scenes specials. Various social media platforms have been used for Olympic content, including Twitter and Facebook. Some marketers credit social media for prompting the official U.S. broadcasters, NBC, to live stream events, including early rounds. == Background == The Olympics is able to advertise to its viewers and its host country with the use of data it collects through Social media marketing. Prominent social media platforms include: Twitter, Facebook, Instagram, Tumblr, YouTube, Google, MSN, Yahoo and many more. Campaign Initiatives and Artificial Intelligence technologies have been used to analyze the social media content of users. Information from consumers such as their preferences, demographics, age and locality are all analyzed to gain consumer insight. Campaign initiatives and AI technologies were used for such purposes in the 2010 Vancouver Winter Olympics and are in use currently. Social media marketing of the Olympics is a new phenomena, beginning prior to the 2008 Beijing Olympics == Variations == There are two classifications of social media marketing recognized by the IOC: Officially sanctioned content from rights holders and sponsors that maximizes the use of Olympic content (imagery, hashtag) Unofficial content that is generated by brands that leverage the excitement of the Olympics == 2008 Beijing Summer Olympics == Social media marketing emerged as a phenomenon during the 2008 Beijing Olympics, which progressed as a marketing and an advertising tactic ever since. The Beijing Olympics became the test subject for social media marketing initiatives started by advertising agencies. In 2008, social media marketing began the transition from one-sided communication to mass communication of the Olympic Games. Although social media marketing of the Olympic Games began in 2008, the audience to the Olympics was still primarily reached through television–reaching an audience of 4.3 billion viewers. At the time, the viewers of the Olympic Games through Internet website platforms made up an audience of approximately 390 million individuals. What was the beginning of Olympic social media marketing, was also the beginning of a more globalized experience of the Olympic Games via social media. Twitter, now a prominent social media platform, began in 2006 and grew to three million active users by the beginning of the 2008 Beijing Olympics. Members of Facebook, another prominent social media platform, tracking the Olympic Games grew from approximately one million during the Olympic Games of Athens 2004 to 90 million during the 2008 Beijing Olympics. Social media use, in general, increased by 24 percent between 2007 and 2008–from 63 percent of U.S. adults to 87 percent of U.S. adults. == 2010 Vancouver Winter Olympics == The International Olympic Committee (IOC) deemed The Vancouver Winter Olympics as "the first social media games” based on its fan base through social media platforms. The IOC launched their Facebook page a month before the games began, attracting 1.5 million fans. Shifting to online viewing attracted a younger audience than past Olympic games with over 60 percent of Facebook fans being under 24 years of age. Athletes like Lindsey Vonn and Shaun White reached fans on social media as the platform posted behind-the-scenes coverage on their experiences. The IOC used social media to create competitions between athletes and fans streamed online. Its YouTube channel hosted a “Best of Us” challenge in which the public could compete in games with their favorite athletes, acquiring three million viewers. Photos spread across social media platforms, such as Flickr, which had 11,000 photos posted by 600 photographers, bringing a new perspective to the games. Twitter contributed constant live updates of the competitions. The IOC's Twitter following doubled to 12,000 followers during the Vancouver Olympics, creating a larger viewer population for the games. The IOC created social media guidelines as more athletes and fans got online to interact with the Olympics. Social media was still relatively new as a marketing platform, so these guidelines confused many individuals. == 2012 London Summer Olympics == The London 2012 Olympic Games succeeded in broadcasting, participation and marketing. For the first time, the IOC broadcast the Olympic Games live and on-demand through YouTube, allowing fans to access the Games anytime, anywhere through live streaming. The combination of conventional broadcasting and mobile platforms reached a global audience of 4.8 billion people. Social media soared with Facebook, Twitter and Google+, attracting 4.7 million followers. Athletes shared photographs, interacted online with fans and updated daily, either in person or via an agent. Instagram was established by 2012, making itself a premier photo-sharing platform perfect for athletes to capture their emotions. Lewis Wiltshire, head of sport for Twitter UK said, "Never before have fans had such direct access to their sporting heroes." Social media created conversation on fan opinions regarding athletes, including 962,756 total mentions of Usain Bolt, “Fastest Man in History,” who defended the 100 meter and 200 meter gold medals. Michael Phelps followed with 828,081 total mentions. Olympic sponsors were active on social media; created several campaigns to promote their brands; and inspired viewers with mass participation and personalized events. The Adidas “Take the Stage” Campaign recognized talent around the world, installing a photo booth and inviting the 550 Olympics athletes to take the stage. (IOC Marketing Report 2012). David Beckham surprised fans at the photo booth in Westfield shopping centre, gaining popularity in UK media. Coca-Cola, Acer Inc., McDonald's, Visa Inc. and several others used similar tactics of participation to attract viewers. == 2014 Sochi Winter Olympics == === Channels === The 2014 Winter Olympic Games were held in Sochi, a city in Krasnodar Krai, Russia, establishing the first “social media Olympics” for Russia. The most popular Russian social media and networking service, VK, created an Olympic page, similar to Facebook's. The Olympic VK page has 2.8 million fans and—the most popular official community on the platform. Throughout the games, VK had 54 million Olympic mentions, an average of 1.5 million per day. Numbers grew on other social media pages: more than 2 million fans joined the Olympic Facebook page, 168,101 followed the Olympic Twitter, 150,000 followed the Olympic Instagram and three million visited the Olympic website in February 2014. There were 90,000 total updates on social media by Sochi 2014 Olympians and teams. The United States was the most active country during the games logging 22,598 posts across Facebook, Twitter, and Instagram. === Engagement === With social media there is also hashtags. The most popular hashtag was #sochi2014 with almost 11,000 uses. The next top five hashtags were #wearewinter, #teamusa, #olympics, #goaus and #wirfuerD. Another popular hashtag was #Sochiproblems, depicting local struggles. Photos of the poor state of Sochi on all platforms made the games the number one trending topic one week before the opening ceremony. #SochiFail and #SochiProblems gave multiple reports of the poor living arrangements, incomplete construction, broken elevators, and polluted waters. This was one way that social media provided awareness to its users. === Media Perceptions === Media perceptions varied during the games; the Olympics was viewed as a confrontation between Eastern and Western Civilizations. The LGBT community took a stand against the games. Sponsors for the games including Coca-Cola, Mcdonald's, and P&G protested against Russian authorities and Russian anti-LGBT laws. Many protests took a stand against Russian laws, which created a discussion between human rights advocates. Advocates believed organizations should not promote certain values in western markets while supporting an anti-human rights government in another market. == 2016 Rio Summer Olympics == Social media marketing was an influential tool in the promotion and analysis of the 2016 Rio Olympics. Thomas Bach, President of the International Olympic Committee said that the power of sport demonstrates that diversity and interconnectedness can enlighten us all. With over 25,000+ sources of accredited media covering the games, the 2016 games were the most consumed Olympic games to date. Marketing for the Rio Olympics began in 2013 and ultimately lasted 3 years. There were 26 million visits to Olympic.org, the official website of the Olympic games, and over 7 billion views of official Olympic content on social media. There were o
Read more →
Similarity learning

Similarity learning is an area of supervised machine learning in artificial intelligence. It is closely related to regression and classification, but the goal is to learn a similarity function that measures how similar or related two objects are. It has applications in ranking, in recommendation systems, visual identity tracking, face verification, and speaker verification. == Learning setup == There are four common setups for similarity and metric distance learning. Regression similarity learning In this setup, pairs of objects are given ( x i 1 , x i 2 ) {\displaystyle (x_{i}^{1},x_{i}^{2})} together with a measure of their similarity y i ∈ R {\displaystyle y_{i}\in R} . The goal is to learn a function that approximates f ( x i 1 , x i 2 ) ∼ y i {\displaystyle f(x_{i}^{1},x_{i}^{2})\sim y_{i}} for every new labeled triplet example ( x i 1 , x i 2 , y i ) {\displaystyle (x_{i}^{1},x_{i}^{2},y_{i})} . This is typically achieved by minimizing a regularized loss min W ∑ i l o s s ( w ; x i 1 , x i 2 , y i ) + r e g ( w ) {\displaystyle \min _{W}\sum _{i}loss(w;x_{i}^{1},x_{i}^{2},y_{i})+reg(w)} . Classification similarity learning Given are pairs of similar objects ( x i , x i + ) {\displaystyle (x_{i},x_{i}^{+})} and non similar objects ( x i , x i − ) {\displaystyle (x_{i},x_{i}^{-})} . An equivalent formulation is that every pair ( x i 1 , x i 2 ) {\displaystyle (x_{i}^{1},x_{i}^{2})} is given together with a binary label y i ∈ { 0 , 1 } {\displaystyle y_{i}\in \{0,1\}} that determines if the two objects are similar or not. The goal is again to learn a classifier that can decide if a new pair of objects is similar or not. Ranking similarity learning Given are triplets of objects ( x i , x i + , x i − ) {\displaystyle (x_{i},x_{i}^{+},x_{i}^{-})} whose relative similarity obey a predefined order: x i {\displaystyle x_{i}} is known to be more similar to x i + {\displaystyle x_{i}^{+}} than to x i − {\displaystyle x_{i}^{-}} . The goal is to learn a function f {\displaystyle f} such that for any new triplet of objects ( x , x + , x − ) {\displaystyle (x,x^{+},x^{-})} , it obeys f ( x , x + ) > f ( x , x − ) {\displaystyle f(x,x^{+})>f(x,x^{-})} (contrastive learning). This setup assumes a weaker form of supervision than in regression, because instead of providing an exact measure of similarity, one only has to provide the relative order of similarity. For this reason, ranking-based similarity learning is easier to apply in real large-scale applications. Locality sensitive hashing (LSH) Hashes input items so that similar items map to the same "buckets" in memory with high probability (the number of buckets being much smaller than the universe of possible input items). It is often applied in nearest neighbor search on large-scale high-dimensional data, e.g., image databases, document collections, time-series databases, and genome databases. A common approach for learning similarity is to model the similarity function as a bilinear form. For example, in the case of ranking similarity learning, one aims to learn a matrix W that parametrizes the similarity function f W ( x , z ) = x T W z {\displaystyle f_{W}(x,z)=x^{T}Wz} . When data is abundant, a common approach is to learn a siamese network – a deep network model with parameter sharing. == Metric learning == Similarity learning is closely related to distance metric learning. Metric learning is the task of learning a distance function over objects. A metric or distance function has to obey four axioms: non-negativity, identity of indiscernibles, symmetry and subadditivity (or the triangle inequality). In practice, metric learning algorithms ignore the condition of identity of indiscernibles and learn a pseudo-metric. When the objects x i {\displaystyle x_{i}} are vectors in R d {\displaystyle R^{d}} , then any matrix W {\displaystyle W} in the symmetric positive semi-definite cone S + d {\displaystyle S_{+}^{d}} defines a distance pseudo-metric of the space of x through the form D W ( x 1 , x 2 ) 2 = ( x 1 − x 2 ) ⊤ W ( x 1 − x 2 ) {\displaystyle D_{W}(x_{1},x_{2})^{2}=(x_{1}-x_{2})^{\top }W(x_{1}-x_{2})} . When W {\displaystyle W} is a symmetric positive definite matrix, D W {\displaystyle D_{W}} is a metric. Moreover, as any symmetric positive semi-definite matrix W ∈ S + d {\displaystyle W\in S_{+}^{d}} can be decomposed as W = L ⊤ L {\displaystyle W=L^{\top }L} where L ∈ R e × d {\displaystyle L\in R^{e\times d}} and e ≥ r a n k ( W ) {\displaystyle e\geq rank(W)} , the distance function D W {\displaystyle D_{W}} can be rewritten equivalently D W ( x 1 , x 2 ) 2 = ( x 1 − x 2 ) ⊤ L ⊤ L ( x 1 − x 2 ) = ‖ L ( x 1 − x 2 ) ‖ 2 2 {\displaystyle D_{W}(x_{1},x_{2})^{2}=(x_{1}-x_{2})^{\top }L^{\top }L(x_{1}-x_{2})=\|L(x_{1}-x_{2})\|_{2}^{2}} . The distance D W ( x 1 , x 2 ) 2 = ‖ x 1 ′ − x 2 ′ ‖ 2 2 {\displaystyle D_{W}(x_{1},x_{2})^{2}=\|x_{1}'-x_{2}'\|_{2}^{2}} corresponds to the Euclidean distance between the transformed feature vectors x 1 ′ = L x 1 {\displaystyle x_{1}'=Lx_{1}} and x 2 ′ = L x 2 {\displaystyle x_{2}'=Lx_{2}} . Many formulations for metric learning have been proposed. Some well-known approaches for metric learning include learning from relative comparisons, which is based on the triplet loss, large margin nearest neighbor, and information theoretic metric learning (ITML). In statistics, the covariance matrix of the data is sometimes used to define a distance metric called Mahalanobis distance. == Applications == Similarity learning is used in information retrieval for learning to rank, in face verification or face identification, and in recommendation systems. Also, many machine learning approaches rely on some metric. This includes unsupervised learning such as clustering, which groups together close or similar objects. It also includes supervised approaches like K-nearest neighbor algorithm which rely on labels of nearby objects to decide on the label of a new object. Metric learning has been proposed as a preprocessing step for many of these approaches. == Scalability == Metric and similarity learning scale quadratically with the dimension of the input space, as can easily see when the learned metric has a bilinear form f W ( x , z ) = x T W z {\displaystyle f_{W}(x,z)=x^{T}Wz} . Scaling to higher dimensions can be achieved by enforcing a sparseness structure over the matrix model, as done with HDSL, and with COMET. == Software == metric-learn is a free software Python library which offers efficient implementations of several supervised and weakly-supervised similarity and metric learning algorithms. The API of metric-learn is compatible with scikit-learn. OpenMetricLearning is a Python framework to train and validate the models producing high-quality embeddings. == Further information == For further information on this topic, see the surveys on metric and similarity learning by Bellet et al. and Kulis.
Read more →
Out-of-band control

Out-of-band control is a method used by network protocols for sending control information (commands, logins, or session signals) separately from the main data, improving reliability and preventing interference. File Transfer Protocol (FTP) employs an out-of-band approach, using one connection for control commands, like logging in or requesting files, and a separate connection for transferring the files themselves.
Read more →
Vans challenge

The Vans challenge is a viral internet challenge that began in March 2019 where people show their Vans shoes landing right-side up after tossing them in the air. The viral sensation reportedly started after a Twitter user shared a video of the occurrence, which was captioned: “Did you know it doesn’t matter how you throw your Vans they will land facing up.” Since then, multiple people on social media posted similar videos of them throwing their Vans in the air and landing right-side up, along with Crocs, UGG boots, and other popular shoes. This theory proved false, as these shoes have not always landed facing up after tossing them.
Read more →