AI App Q

AI App Q — independent reviews, comparisons, pricing and step-by-step guides on Aizhi.

• 

  Azure Stream Analytics

  Microsoft Azure Stream Analytics is a serverless scalable complex event processing engine by Microsoft that enables users to develop and run real-time analytics on multiple streams of data from sources such as devices, sensors, web sites, social media, and other applications. Users can set up alerts to detect anomalies, predict trends, trigger necessary workflows when certain conditions are observed, and make data available to other downstream applications and services for presentation, archiving, or further analysis. == Query Language == Users can author real-time analytics using a simple declarative SQL-like language with embedded support for temporal logic. Callouts to custom code with JavaScript user defined functions extend the streaming logic written in SQL. Callouts to Azure Machine Learning helps with predictive scoring on streaming data. == Scalability == Azure Stream Analytics is a serverless job service on Azure that eliminates the need for infrastructure, servers, virtual machines, or managed clusters. Users only pay for the processing used for the running jobs. == IoT applications == Azure Stream Analytics integrates with Azure IoT Hub to enable real-time analytics on data from IoT devices and applications. == Real-time Dashboards == Users can build real-time dashboards with Power BI for a live command and control view. Real-time dashboards help transform live data into actionable and insightful visuals. == Data Input Sources == Stream Analytics supports three different types of input sources - Azure Event Hubs, Azure IoT Hubs, and Azure Blob Storage. Additionally, stream analytics supports Azure Blob storage as the input reference data to help augment fast moving event data streams with static data. Stream analytics supports a wide variety of output targets. Support for Power BI allows for real-time dashboarding. Event Hub, Service bus topics and queues help trigger downstream workflows. Support for Azure Table Storage, Azure SQL Databases, Azure SQL Data Warehouse, Azure SQL, Document DB, Azure Data Lake Store enable a variety of downstream analysis and archiving capabilities.

  Read more →
• 

  Knights of Sidonia

  Knights of Sidonia (Japanese: シドニアの騎士, Hepburn: Shidonia no Kishi) is a Japanese manga series written and illustrated by Tsutomu Nihei. It was serialized by Kodansha's seinen manga magazine Monthly Afternoon between April 2009 and September 2015, with its chapters collected in 15 tankōbon volumes. It tells the story of Nagate Tanikaze, an "under-dweller" destined to become a Garde pilot, whose mission is to defend the generation ship Sidonia from a hostile alien species called Gauna. The manga was licensed for English release in North America by Vertical. An anime television series adaptation was produced by Polygon Pictures. The first season aired from April to June 2014; the second between April and June 2015. An anime film sequel titled Knights of Sidonia: Love Woven in the Stars premiered in June 2021. In 2015, Knights of Sidonia received the 39th Kodansha Manga Award in the general category, as well as the 47th Seiun Award in the Best Comic category in 2016. == Plot == === Setting === The story is set in the year 3394, a thousand years after mankind flees from Earth after it was destroyed by a race of shapeshifting aliens called the Gauna (奇居子（ガウナ）), aboard hundreds of colossal spacecraft created from the remains of the planet. One such ship is the Sidonia, which has developed its own human culture closely based on that of Japan where human cloning, asexual reproduction, and human genetic engineering, such as granting humans photosynthesis, are commonplace. It is also revealed that the top echelons of this society have secretly been granted immortality. With a population of over 500,000 people, Sidonia is possibly the last human settlement remaining, as the fates of the other ships are unknown. Little is known about the true nature of the Gauna or their motivation for attacking humanity. At any given time, a Gauna consists of a nearly impenetrable core protected by a dense layer of malleable flesh known as "placenta" (胞衣, ena). Once the ena is shed away and the core is destroyed, the Gauna's body disintegrates. While Sidonia itself is heavily armed with an arsenal of high-output beam cannons and mass cannons including slow but powerful planet-destroying warheads, it is primarily defended by large mechanized weapons called Gardes (衛人, Morito) whose weaponry and mobility is powered by "Higgs particles" (ヘイグス粒子, Heigusu Ryūshi), armed with a high-output beam cannon for long range assaults and a special spear known as "Kabizashi" for close combat. The tip of the kabizashi is made of a rare and little-understood material which has the unique property of being able to destroy a Gauna's core. Later the Gardes are also equipped with firearms whose ammunition have the same material of the Kabizashi after a means to artificially mass-produce it is discovered. Most people in the surviving human population are screened and drafted as Garde pilots at a young age, if they are shown to be capable of piloting them. === Story === The story follows the adventures of Garde pilot Nagate Tanikaze, who lived in the underground layer of Sidonia since birth and was raised by his grandfather. Never having met anyone else, he trains himself in an old Guardian pilot simulator every day, eventually mastering it. After his grandfather's death, he emerges to the surface and is selected as a Garde pilot, just as Sidonia is once again threatened by the Gauna. == Media == === Manga === Written and illustrated by Tsutomu Nihei, Knights of Sidonia was serialized in Kodansha's seinen manga magazine Monthly Afternoon from April 25, 2009, to September 25, 2015. It was compiled in 15 tankōbon volumes. The manga has been licensed in North America by Vertical, who released all 15 volumes in English between February 5, 2013, and April 26, 2016. === Anime === An anime television series adaptation, produced by Polygon Pictures, aired its first season from April 10 to June 26, 2014, on MBS and later on TBS, CBC and BS-TBS. The series was directed by Kōbun Shizuno, assisted by Hiroyuki Seshita, with scripts by Sadayuki Murai and character designs by Yuki Moriyama. The opening theme song is "Sidonia" (シドニア, Shidonia), performed by Angela, while the ending theme song is "Show" (掌 -show-, Shō), performed by Eri Kitamura. A second season aired from April 11 to June 26, 2015. For the second season, the opening theme song is "Kishi Kōshinkyoku" (騎士行進曲, Knight March), performed by Angela, while the ending theme song is "Requiem" (鎮魂歌 -レクイエム-, Rekuiemu), performed by CustomiZ. The series was localized and streamed by Netflix in all of its territories since July 4, 2014, becoming the service's first original anime, as well as the first anime series on Netflix available in Dolby Vision/HDR. The first season has been licensed for home video release by Sentai Filmworks. The second season was released on Netflix on July 3, 2015, and has been licensed by Sentai Filmworks for home video distribution. In July 2021, Funimation announced they acquired the streaming rights from Netflix to both seasons. === Films === A compilation film of the first season with additional scenes and re-edited sound effects was released on March 6, 2015. A new anime film, titled Knights of Sidonia: Love Woven in the Stars, was announced on July 3, 2020. Hiroyuki Seshita served as chief director, while Tadahiro Yoshihira served as director for the new film, with Polygon Pictures returning for production. Sadayuki Murai and Tetsuya Yamada returned to write scripts, while Shūji Katayama composed the music. The rest of the staff and cast returned to reprise their roles. The first four minutes of the film were shown on YouTube on April 28, 2021. The film was set to premiere on May 14, 2021, but was delayed to June 4, 2021, due to the COVID-19 pandemic. Funimation screened the film in international theaters starting on September 13, 2021. == Reception == === Manga === Knights of Sidonia won the 39th Kodansha Manga Award in the general category in 2015. The manga won the 47th Seiun Award in the Best Comic category in 2016. It also won the Best Seinen category at the 26th Salón del Manga de Barcelona in 2020. It was one of the Jury Recommended works in the Manga Division at the 17th Japan Media Arts Festival in 2013. The Young Adult Library Services Association listed Knights of Sidonia in its 2014 list of Top 10 Graphic Novels for Teens. Carlo Santos from Anime News Network gave the first manga volume a B, stating, "It is got a young man piloting a giant robot against alien enemies, but Knight of Sidonia is no Neon Genesis Evangelion. Yet it is not as bleak or incomprehensible as Tsutomu Nihei works like Blame! or Biomega, either—rather, it is the best of both worlds, bringing Nihei's hard sci-fi mentality into a more conventional space-adventure environment". === Anime === The anime series received positive reviews, even from famous members of the Japanese anime/game industry, like Hideo Kojima, creator of the Metal Gear series, who claims that "It's a kind of anime that we haven't seen for a while that has that sci-fi spirit. Using digital technology cultivated through games, it creates animation that encapsulates Japan's cultural assets like manga, cel animation, kanji, giant robots, etc. What's born is a unique made-in-Japan work that could never be cooked up in Hollywood. Japanese culture has lost its 'cool', and Knights of Sidonia will be the white knight that saves it". Other industry pros left acknowledgements as well, including Akiko Higashimura, Digitarou and Yoshinao Dao.

  Read more →
• 

  Anytime algorithm

  In computer science, an anytime algorithm is an algorithm that can return a valid solution to a problem even if it is interrupted before it ends. The algorithm is expected to find better and better solutions the longer it keeps running. Most algorithms run to completion: they provide a single answer after performing some fixed amount of computation. In some cases, however, the user may wish to terminate the algorithm prior to completion. The amount of computation required may be substantial, for example, and computational resources might need to be reallocated. Most algorithms either run to completion or they provide no useful solution information. Anytime algorithms, however, are able to return a partial answer, whose quality depends on the amount of computation they were able to perform. The answer generated by anytime algorithms is an approximation of the correct answer. == Names == An anytime algorithm may be also called an "interruptible algorithm". They are different from contract algorithms, which must declare a time in advance; in an anytime algorithm, a process can just announce that it is terminating. == Goals == The goal of anytime algorithms are to give intelligent systems the ability to make results of better quality in return for turn-around time. They are also supposed to be flexible in time and resources. They are important because artificial intelligence or AI algorithms can take a long time to complete results. This algorithm is designed to complete in a shorter amount of time. Also, these are intended to have a better understanding that the system is dependent and restricted to its agents and how they work cooperatively. An example is the Newton–Raphson iteration applied to finding the square root of a number. Another example that uses anytime algorithms is trajectory problems when you're aiming for a target; the object is moving through space while waiting for the algorithm to finish and even an approximate answer can significantly improve its accuracy if given early. What makes anytime algorithms unique is their ability to return many possible outcomes for any given input. An anytime algorithm uses many well defined quality measures to monitor progress in problem solving and distributed computing resources. It keeps searching for the best possible answer with the amount of time that it is given. It may not run until completion and may improve the answer if it is allowed to run longer. This is often used for large decision set problems. This would generally not provide useful information unless it is allowed to finish. While this may sound similar to dynamic programming, the difference is that it is fine-tuned through random adjustments, rather than sequential. Anytime algorithms are designed so that it can be told to stop at any time and would return the best result it has found so far. This is why it is called an interruptible algorithm. Certain anytime algorithms also maintain the last result, so that if they are given more time, they can continue from where they left off to obtain an even better result. == Decision trees == When the decider has to act, there must be some ambiguity. Also, there must be some idea about how to solve this ambiguity. This idea must be translatable to a state to action diagram. == Performance profile == The performance profile estimates the quality of the results based on the input and the amount of time that is allotted to the algorithm. The better the estimate, the sooner the result would be found. Some systems have a larger database that gives the probability that the output is the expected output. One algorithm can have several performance profiles. Most of the time performance profiles are constructed using mathematical statistics using representative cases. For example, in the traveling salesman problem, the performance profile was generated using a user-defined special program to generate the necessary statistics. In this example, the performance profile is the mapping of time to the expected results. This quality can be measured in several ways: certainty: where probability of correctness determines quality accuracy: where error bound determines quality specificity: where the amount of particulars determine quality == Algorithm prerequisites == Initial behavior: While some algorithms start with immediate guesses, others take a more calculated approach and have a start up period before making any guesses. Growth direction: How the quality of the program's "output" or result, varies as a function of the amount of time ("run time") Growth rate: Amount of increase with each step. Does it change constantly, such as in a bubble sort or does it change unpredictably? End condition: The amount of runtime needed

  Read more →
• 

  Batch normalization

  In artificial neural networks, batch normalization (also known as batch norm) is a normalization technique used to make training faster and more stable by adjusting the inputs to each layer—re-centering them around zero and re-scaling them to a standard size. It was introduced by Sergey Ioffe and Christian Szegedy in 2015. Experts still debate why batch normalization works so well. It was initially thought to tackle internal covariate shift, a problem where parameter initialization and changes in the distribution of the inputs of each layer affect the learning rate of the network. However, newer research suggests it doesn’t fix this shift but instead smooths the objective function—a mathematical guide the network follows to improve—enhancing performance. In very deep networks, batch normalization can initially cause a severe gradient explosion—where updates to the network grow uncontrollably large—but this is managed with shortcuts called skip connections in residual networks. Another theory is that batch normalization adjusts data by handling its size and path separately, speeding up training. == Internal covariate shift == Each layer in a neural network has inputs that follow a specific distribution, which shifts during training due to two main factors: the random starting values of the network’s settings (parameter initialization) and the natural variation in the input data. This shifting pattern affecting the inputs to the network’s inner layers is called internal covariate shift. While a strict definition isn’t fully agreed upon, experiments show that it involves changes in the means and variances of these inputs during training. Batch normalization was first developed to address internal covariate shift. During training, as the parameters of preceding layers adjust, the distribution of inputs to the current layer changes accordingly, such that the current layer needs to constantly readjust to new distributions. This issue is particularly severe in deep networks, because small changes in shallower hidden layers will be amplified as they propagate within the network, resulting in significant shift in deeper hidden layers. Batch normalization was proposed to reduced these unwanted shifts to speed up training and produce more reliable models. Beyond possibly tackling internal covariate shift, batch normalization offers several additional advantages. It allows the network to use a higher learning rate—a setting that controls how quickly the network learns—without causing problems like vanishing or exploding gradients, where updates become too small or too large. It also appears to have a regularizing effect, improving the network’s ability to generalize to new data, reducing the need for dropout, a technique used to prevent overfitting (when a model learns the training data too well and fails on new data). Additionally, networks using batch normalization are less sensitive to the choice of starting settings or learning rates, making them more robust and adaptable. == Procedures == === Transformation === In a neural network, batch normalization is achieved through a normalization step that fixes the means and variances of each layer's inputs. Ideally, the normalization would be conducted over the entire training set, but to use this step jointly with stochastic optimization methods, it is impractical to use the global information. Thus, normalization is restrained to each mini-batch in the training process. Let us use B to denote a mini-batch of size m of the entire training set. The empirical mean and variance of B could thus be denoted as μ B = 1 m ∑ i = 1 m x i {\displaystyle \mu _{B}={\frac {1}{m}}\sum _{i=1}^{m}x_{i}} and σ B 2 = 1 m ∑ i = 1 m ( x i − μ B ) 2 {\displaystyle \sigma _{B}^{2}={\frac {1}{m}}\sum _{i=1}^{m}(x_{i}-\mu _{B})^{2}} . For a layer of the network with d-dimensional input, x = ( x ( 1 ) , . . . , x ( d ) ) {\displaystyle x=(x^{(1)},...,x^{(d)})} , each dimension of its input is then normalized (i.e. re-centered and re-scaled) separately, x ^ i ( k ) = x i ( k ) − μ B ( k ) ( σ B ( k ) ) 2 + ϵ {\displaystyle {\hat {x}}_{i}^{(k)}={\frac {x_{i}^{(k)}-\mu _{B}^{(k)}}{\sqrt {\left(\sigma _{B}^{(k)}\right)^{2}+\epsilon }}}} , where k ∈ [ 1 , d ] {\displaystyle k\in [1,d]} and i ∈ [ 1 , m ] {\displaystyle i\in [1,m]} ; μ B ( k ) {\displaystyle \mu _{B}^{(k)}} and σ B ( k ) {\displaystyle \sigma _{B}^{(k)}} are the per-dimension mean and standard deviation, respectively. ϵ {\displaystyle \epsilon } is added in the denominator for numerical stability and is an arbitrarily small positive constant. The resulting normalized activation x ^ ( k ) {\displaystyle {\hat {x}}^{(k)}} have zero mean and unit variance, if ϵ {\displaystyle \epsilon } is not taken into account. To restore the representation power of the network, a transformation step then follows as y i ( k ) = γ ( k ) x ^ i ( k ) + β ( k ) {\displaystyle y_{i}^{(k)}=\gamma ^{(k)}{\hat {x}}_{i}^{(k)}+\beta ^{(k)}} , where the parameters γ ( k ) {\displaystyle \gamma ^{(k)}} and β ( k ) {\displaystyle \beta ^{(k)}} are subsequently learned in the optimization process. Formally, the operation that implements batch normalization is a transform B N γ ( k ) , β ( k ) : x 1... m ( k ) → y 1... m ( k ) {\displaystyle BN_{\gamma ^{(k)},\beta ^{(k)}}:x_{1...m}^{(k)}\rightarrow y_{1...m}^{(k)}} called the Batch Normalizing transform. The output of the BN transform y ( k ) = B N γ ( k ) , β ( k ) ( x ( k ) ) {\displaystyle y^{(k)}=BN_{\gamma ^{(k)},\beta ^{(k)}}(x^{(k)})} is then passed to other network layers, while the normalized output x ^ i ( k ) {\displaystyle {\hat {x}}_{i}^{(k)}} remains internal to the current layer. === Backpropagation === The described BN transform is a differentiable operation, and the gradient of the loss l {\displaystyle l} with respect to the different parameters can be computed directly with the chain rule. Specifically, ∂ l ∂ y i ( k ) {\displaystyle {\frac {\partial l}{\partial y_{i}^{(k)}}}} depends on the choice of activation function, and the gradient against other parameters could be expressed as a function of ∂ l ∂ y i ( k ) {\displaystyle {\frac {\partial l}{\partial y_{i}^{(k)}}}} : ∂ l ∂ x ^ i ( k ) = ∂ l ∂ y i ( k ) γ ( k ) {\displaystyle {\frac {\partial l}{\partial {\hat {x}}_{i}^{(k)}}}={\frac {\partial l}{\partial y_{i}^{(k)}}}\gamma ^{(k)}} , ∂ l ∂ γ ( k ) = ∑ i = 1 m ∂ l ∂ y i ( k ) x ^ i ( k ) {\displaystyle {\frac {\partial l}{\partial \gamma ^{(k)}}}=\sum _{i=1}^{m}{\frac {\partial l}{\partial y_{i}^{(k)}}}{\hat {x}}_{i}^{(k)}} , ∂ l ∂ β ( k ) = ∑ i = 1 m ∂ l ∂ y i ( k ) {\displaystyle {\frac {\partial l}{\partial \beta ^{(k)}}}=\sum _{i=1}^{m}{\frac {\partial l}{\partial y_{i}^{(k)}}}} , ∂ l ∂ σ B ( k ) 2 = ∑ i = 1 m ∂ l ∂ y i ( k ) ( x i ( k ) − μ B ( k ) ) ( − γ ( k ) 2 ( σ B ( k ) 2 + ϵ ) − 3 / 2 ) {\displaystyle {\frac {\partial l}{\partial \sigma _{B}^{(k)^{2}}}}=\sum _{i=1}^{m}{\frac {\partial l}{\partial y_{i}^{(k)}}}(x_{i}^{(k)}-\mu _{B}^{(k)})\left(-{\frac {\gamma ^{(k)}}{2}}(\sigma _{B}^{(k)^{2}}+\epsilon )^{-3/2}\right)} , ∂ l ∂ μ B ( k ) = ∑ i = 1 m ∂ l ∂ y i ( k ) − γ ( k ) σ B ( k ) 2 + ϵ + ∂ l ∂ σ B ( k ) 2 1 m ∑ i = 1 m ( − 2 ) ⋅ ( x i ( k ) − μ B ( k ) ) {\displaystyle {\frac {\partial l}{\partial \mu _{B}^{(k)}}}=\sum _{i=1}^{m}{\frac {\partial l}{\partial y_{i}^{(k)}}}{\frac {-\gamma ^{(k)}}{\sqrt {\sigma _{B}^{(k)^{2}}+\epsilon }}}+{\frac {\partial l}{\partial \sigma _{B}^{(k)^{2}}}}{\frac {1}{m}}\sum _{i=1}^{m}(-2)\cdot (x_{i}^{(k)}-\mu _{B}^{(k)})} , and ∂ l ∂ x i ( k ) = ∂ l ∂ x ^ i ( k ) 1 σ B ( k ) 2 + ϵ + ∂ l ∂ σ B ( k ) 2 2 ( x i ( k ) − μ B ( k ) ) m + ∂ l ∂ μ B ( k ) 1 m {\displaystyle {\frac {\partial l}{\partial x_{i}^{(k)}}}={\frac {\partial l}{\partial {\hat {x}}_{i}^{(k)}}}{\frac {1}{\sqrt {\sigma _{B}^{(k)^{2}}+\epsilon }}}+{\frac {\partial l}{\partial \sigma _{B}^{(k)^{2}}}}{\frac {2(x_{i}^{(k)}-\mu _{B}^{(k)})}{m}}+{\frac {\partial l}{\partial \mu _{B}^{(k)}}}{\frac {1}{m}}} . === Inference === During the training stage, the normalization steps depend on the mini-batches to ensure efficient and reliable training. However, in the inference stage, this dependence is not useful any more. Instead, the normalization step in this stage is computed with the population statistics such that the output could depend on the input in a deterministic manner. The population mean, E [ x ( k ) ] {\displaystyle E[x^{(k)}]} , and variance, Var ⁡ [ x ( k ) ] {\displaystyle \operatorname {Var} [x^{(k)}]} , are computed as: E [ x ( k ) ] = E B [ μ B ( k ) ] {\displaystyle E[x^{(k)}]=E_{B}[\mu _{B}^{(k)}]} , and Var ⁡ [ x ( k ) ] = m m − 1 E B [ ( σ B ( k ) ) 2 ] {\displaystyle \operatorname {Var} [x^{(k)}]={\frac {m}{m-1}}E_{B}[\left(\sigma _{B}^{(k)}\right)^{2}]} . The population statistics thus is a complete representation of the mini-batches. The BN transform in the inference step thus becomes y ( k ) = B N γ ( k ) , β ( k ) inf ( x ( k ) ) = γ ( k ) x ( k ) − E [ x ( k ) ] Var ⁡ [ x ( k ) ] + ϵ + β

  Read more →
• 

  Cloem

  Cloem is a company based in Cannes, France, which applies natural language processing (NLP) technologies to assist patent applicants in creating variants of patent claims, called "cloems". According to the company, these "computer-generated claims can be published to keep potential competitors from attempting to file adjacent patent claims." == Technology == According to Cloem, dictionaries, ontologies and proprietary claim-drafting algorithms are used to draft alternative claims based on a client's original set of claims. In particular, the original set of claims is subject to various permutations and linguistic manipulations "by considering alternative definitions for terms as well as “synonyms, hyponyms, hyperonyms, meronyms, holonyms, and antonyms.”" == Possible uses == Cloem can optionally publish one or more created texts, as electronic publications or as paper-printed publications. These can potentially serve – through a defensive publication – as prior art to prevent another party for obtaining a patent on the subject-matter at stake. In other words, after an initial patent filing, an "improvement" patent (adjacent invention) can be applied for by another party, such as a competitor. By publishing variants of a patent claim, the risk of adverse patenting may potentially be decreased (improvement inventions may no longer be patentable). Cloems may also be potentially patentable. One of the issues of patentability, however, is that only a natural person can be a listed as an inventor on a patent. Since cloems are produced by a computer based on a person's input, it is not clear if the computer or the person is the inventor. The inventorship of Cloem texts is an open question.

  Read more →
• 

  Tamarin Prover

  Tamarin Prover is a computer software program for formal verification of cryptographic protocols. It has been used to verify Transport Layer Security 1.3, ISO/IEC 9798, DNP3 Secure Authentication v5, WireGuard, and the PQ3 Messaging Protocol of Apple iMessage. Tamarin is an open source tool, written in Haskell, built as a successor to an older verification tool called Scyther. Tamarin has automatic proof features, but can also be self-guided. In Tamarin lemmas that representing security properties are defined. After changes are made to a protocol, Tamarin can verify if the security properties are maintained. The results of a Tamarin execution will either be a proof that the security property holds within the protocol, an example protocol run where the security property does not hold, or Tamarin could potentially fail to halt.

  Read more →
• 

  Conceptual dependency theory

  Conceptual dependency theory is a model of natural language understanding used in artificial intelligence systems. Roger Schank at Stanford University introduced the model in 1969, in the early days of artificial intelligence. This model was extensively used by Schank's students at Yale University such as Robert Wilensky, Wendy Lehnert, and Janet Kolodner. Schank developed the model to represent knowledge for natural language input into computers. Partly influenced by the work of Sydney Lamb, his goal was to make the meaning independent of the words used in the input, i.e. two sentences identical in meaning would have a single representation. The system was also intended to draw logical inferences. The model uses the following basic representational tokens: real world objects, each with some attributes. real world actions, each with attributes times locations A set of conceptual transitions then act on this representation, e.g. an ATRANS is used to represent a transfer such as "give" or "take" while a PTRANS is used to act on locations such as "move" or "go". An MTRANS represents mental acts such as "tell", etc. A sentence such as "John gave a book to Mary" is then represented as the action of an ATRANS on two real world objects, John and Mary.

  Read more →
• 

  Neural architecture search

  Neural architecture search (NAS) is a technique for automating the design of artificial neural networks (ANN), a widely used model in the field of machine learning. NAS has been used to design networks that are on par with or outperform hand-designed architectures. Methods for NAS can be categorized according to the search space, search strategy and performance estimation strategy used: The search space defines the type(s) of ANN that can be designed and optimized. The search strategy defines the approach used to explore the search space. The performance estimation strategy evaluates the performance of a possible ANN from its design (without constructing and training it). NAS is closely related to hyperparameter optimization and meta-learning and is a subfield of automated machine learning (AutoML). == Reinforcement learning == Reinforcement learning (RL) can underpin a NAS search strategy. Barret Zoph and Quoc Viet Le applied NAS with RL targeting the CIFAR-10 dataset and achieved a network architecture that rivals the best manually-designed architecture for accuracy, with an error rate of 3.65, 0.09 percent better and 1.05x faster than a related hand-designed model. On the Penn Treebank dataset, that model composed a recurrent cell that outperforms LSTM, reaching a test set perplexity of 62.4, or 3.6 perplexity better than the prior leading system. On the PTB character language modeling task it achieved bits per character of 1.214. Learning a model architecture directly on a large dataset can be a lengthy process. NASNet addressed this issue by transferring a building block designed for a small dataset to a larger dataset. The design was constrained to use two types of convolutional cells to return feature maps that serve two main functions when convoluting an input feature map: normal cells that return maps of the same extent (height and width) and reduction cells in which the returned feature map height and width is reduced by a factor of two. For the reduction cell, the initial operation applied to the cell's inputs uses a stride of two (to reduce the height and width). The learned aspect of the design included elements such as which lower layer(s) each higher layer took as input, the transformations applied at that layer and to merge multiple outputs at each layer. In the studied example, the best convolutional layer (or "cell") was designed for the CIFAR-10 dataset and then applied to the ImageNet dataset by stacking copies of this cell, each with its own parameters. The approach yielded accuracy of 82.7% top-1 and 96.2% top-5. This exceeded the best human-invented architectures at a cost of 9 billion fewer FLOPS—a reduction of 28%. The system continued to exceed the manually-designed alternative at varying computation levels. The image features learned from image classification can be transferred to other computer vision problems. E.g., for object detection, the learned cells integrated with the Faster-RCNN framework improved performance by 4.0% on the COCO dataset. In the so-called Efficient Neural Architecture Search (ENAS), a controller discovers architectures by learning to search for an optimal subgraph within a large graph. The controller is trained with policy gradient to select a subgraph that maximizes the validation set's expected reward. The model corresponding to the subgraph is trained to minimize a canonical cross entropy loss. Multiple child models share parameters, ENAS requires fewer GPU-hours than other approaches and 1000-fold less than "standard" NAS. On CIFAR-10, the ENAS design achieved a test error of 2.89%, comparable to NASNet. On Penn Treebank, the ENAS design reached test perplexity of 55.8. == Evolution == An alternative approach to NAS is based on evolutionary algorithms, which has been employed by several groups. An Evolutionary Algorithm for Neural Architecture Search generally performs the following procedure. First a pool consisting of different candidate architectures along with their validation scores (fitness) is initialised. At each step the architectures in the candidate pool are mutated (e.g.: 3x3 convolution instead of a 5x5 convolution). Next the new architectures are trained from scratch for a few epochs and their validation scores are obtained. This is followed by replacing the lowest scoring architectures in the candidate pool with the better, newer architectures. This procedure is repeated multiple times and thus the candidate pool is refined over time. Mutations in the context of evolving ANNs are operations such as adding or removing a layer, which include changing the type of a layer (e.g., from convolution to pooling), changing the hyperparameters of a layer, or changing the training hyperparameters. On CIFAR-10 and ImageNet, evolution and RL performed comparably, while both slightly outperformed random search. == Bayesian optimization == Bayesian Optimization (BO), which has proven to be an efficient method for hyperparameter optimization, can also be applied to NAS. In this context, the objective function maps an architecture to its validation error after being trained for a number of epochs. At each iteration, BO uses a surrogate to model this objective function based on previously obtained architectures and their validation errors. One then chooses the next architecture to evaluate by maximizing an acquisition function, such as expected improvement, which provides a balance between exploration and exploitation. Acquisition function maximization and objective function evaluation are often computationally expensive for NAS, and make the application of BO challenging in this context. Recently, BANANAS has achieved promising results in this direction by introducing a high-performing instantiation of BO coupled to a neural predictor. == Hill-climbing == Another group used a hill climbing procedure that applies network morphisms, followed by short cosine-annealing optimization runs. The approach yielded competitive results, requiring resources on the same order of magnitude as training a single network. E.g., on CIFAR-10, the method designed and trained a network with an error rate below 5% in 12 hours on a single GPU. == Multi-objective search == While most approaches solely focus on finding architecture with maximal predictive performance, for most practical applications other objectives are relevant, such as memory consumption, model size or inference time (i.e., the time required to obtain a prediction). Because of that, researchers created a multi-objective search. LEMONADE is an evolutionary algorithm that adopted Lamarckism to efficiently optimize multiple objectives. In every generation, child networks are generated to improve the Pareto frontier with respect to the current population of ANNs. Neural Architect is claimed to be a resource-aware multi-objective RL-based NAS with network embedding and performance prediction. Network embedding encodes an existing network to a trainable embedding vector. Based on the embedding, a controller network generates transformations of the target network. A multi-objective reward function considers network accuracy, computational resource and training time. The reward is predicted by multiple performance simulation networks that are pre-trained or co-trained with the controller network. The controller network is trained via policy gradient. Following a modification, the resulting candidate network is evaluated by both an accuracy network and a training time network. The results are combined by a reward engine that passes its output back to the controller network. == One-shot models == RL or evolution-based NAS require thousands of GPU-days of searching/training to achieve state-of-the-art computer vision results as described in the NASNet, mNASNet and MobileNetV3 papers. To reduce computational cost, many recent NAS methods rely on the weight-sharing idea. In this approach, a single overparameterized supernetwork (also known as the one-shot model) is defined. A supernetwork is a very large Directed Acyclic Graph (DAG) whose subgraphs are different candidate neural networks. Thus, in a supernetwork, the weights are shared among a large number of different sub-architectures that have edges in common, each of which is considered as a path within the supernet. The essential idea is to train one supernetwork that spans many options for the final design rather than generating and training thousands of networks independently. In addition to the learned parameters, a set of architecture parameters are learnt to depict preference for one module over another. Such methods reduce the required computational resources to only a few GPU days. More recent works further combine this weight-sharing paradigm, with a continuous relaxation of the search space, which enables the use of gradient-based optimization methods. These approaches are generally referred to as differentiable NAS and have proven very efficient in exploring the search space of ne

  Read more →
• 

  Data item

  A data item describes an atomic state of a particular object concerning a specific property at a certain time point. A collection of data items for the same object at the same time forms an object instance (or table row). Any type of complex information can be broken down to elementary data items (atomic state). Data items are identified by object (o), property (p) and time (t), while the value (v) is a function of o, p and t: v = F(o,p,t). Values typically are represented by symbols like numbers, texts, images, sounds or videos. Values are not necessarily atomic. A value's complexity depends on the complexity of the property and time component. When looking at databases or XML files, the object is usually identified by an object name or other type of object identifier, which is part of the "data". Properties are defined as columns (table row), properties (object instance) or tags (XML). Often, time is not explicitly expressed and is an attribute applying to the complete data set. Other data collections provide time on the instance level (time series), column level, or even attribute/property level.

  Read more →
• 

  Sinewave synthesis

  Sinewave synthesis, or sine wave speech, is a technique for synthesizing speech by replacing the formants (main bands of energy) with pure tone whistles. The first sinewave synthesis program (SWS) for the automatic creation of stimuli for perceptual experiments was developed by Philip Rubin at Haskins Laboratories in the 1970s. This program was subsequently used by Robert Remez, Philip Rubin, David Pisoni, and other colleagues to show that listeners can perceive continuous speech without traditional speech cues, i.e., pitch, stress, and intonation. This work paved the way for a view of speech as a dynamic pattern of trajectories through articulatory-acoustic space.

  Read more →
• 

  Fuzzy mathematics

  Fuzzy mathematics is a branch of mathematics that extends classical set theory and logic to model reasoning under uncertainty. Initiated by Lotfi Asker Zadeh in 1965 with the introduction of fuzzy sets, the field has since evolved to include fuzzy set theory, fuzzy logic, and various fuzzy analogues of traditional mathematic structures. Unlike classical mathematics, which usually relies on binary membership (an element either belongs to a set or it does not), fuzzy mathematics allows elements to partially belong to a set, with degrees of membership represented by values in the interval [0, 1]. This framework enables more flexible modeling of imprecise or vague concepts. Fuzzy mathematics has found applications in numerous domains, including control theory, artificial intelligence, decision theory, pattern recognition, and linguistics, where the modeling of gradations and uncertainty is essential. == Definition == A fuzzy subset A of a set X is defined by a function A: X → L, where L is typically the interval [0, 1]. This function is called the membership function of the fuzzy subset and assigns to each element x in X a degree of membership A(x) in the fuzzy set A. In classical set theory, a subset of X can be represented by an indicator function (also known as a characteristic function), which maps elements to either 0 or 1, indicating non-membership or full membership, respectively. Fuzzy subsets generalize this concept by allowing any real value between 0 and 1, thereby enabling partial membership. More generally, the codomain L of the membership function can be replaced with any complete lattice, resulting in the broader framework of L-fuzzy sets. == Fuzzification == The development of fuzzification in mathematics can be broadly divided into three historical stages: Initial, straightforward fuzzifications (1960s–1970s), Expansion of generalization techniques (1980s), Standardization, axiomatization, and L-fuzzification (1990s). Fuzzification generally involves extending classical mathematical concepts from binary (crisp) logic, where membership is determined by characteristic functions, to fuzzy logic, where membership is expressed by values in the interval [0, 1] via membership functions. Let A and B be fuzzy subsets of a set X. The fuzzy versions of set-theoretic operations are commonly defined as: ( A ∩ B ) ( x ) = min ( A ( x ) , B ( x ) ) {\displaystyle (A\cap B)(x)=\min(A(x),B(x))} ( A ∪ B ) ( x ) = max ( A ( x ) , B ( x ) ) {\displaystyle (A\cup B)(x)=\max(A(x),B(x))} for all x ∈ X {\displaystyle x\in X} . These operations can be generalized using t-norms and t-conorms, respectively. For example, the minimum operation can be replaced by multiplication: ( A ∩ B ) ( x ) = A ( x ) ⋅ B ( x ) {\displaystyle (A\cap B)(x)=A(x)\cdot B(x)} Fuzzification of algebraic structures often relies on generalizing the closure property. Let ∗ {\displaystyle } be a binary operation on X, and let A be a fuzzy subset of X. Then A is said to satisfy fuzzy closure if: A ( x ∗ y ) ≥ min ( A ( x ) , A ( y ) ) {\displaystyle A(xy)\geq \min(A(x),A(y))} for all x , y ∈ X {\displaystyle x,y\in X} . If ( G , ∗ ) {\displaystyle (G,)} is a group, then a fuzzy subset A of G is a fuzzy subgroup if: A ( x ∗ y − 1 ) ≥ min ( A ( x ) , A ( y − 1 ) ) {\displaystyle A(xy^{-1})\geq \min(A(x),A(y^{-1}))} for all x , y ∈ G {\displaystyle x,y\in G} . Similar generalizations apply to relational properties. For example, for example, for fuzzification of the transitivity property, a fuzzy relation R {\displaystyle R} on X {\displaystyle X} (i.e., a fuzzy subset of X × X {\displaystyle X\times X} ) is said to be fuzzy transitive if: R ( x , z ) ≥ min ( R ( x , y ) , R ( y , z ) ) {\displaystyle R(x,z)\geq \min(R(x,y),R(y,z))} for all x , y , z ∈ X {\displaystyle x,y,z\in X} . == Fuzzy analogues == Fuzzy subgroupoids and fuzzy subgroups were introduced in 1971 by A. Rosenfeld. Analogues of other mathematical subjects have been translated to fuzzy mathematics, such as fuzzy field theory and fuzzy Galois theory, fuzzy topology, fuzzy geometry, fuzzy orderings, and fuzzy graphs.

  Read more →
• 

  Fuzzy control system

  A fuzzy control system is a control system based on fuzzy logic – a mathematical system that analyzes analog input values in terms of logical variables that take on continuous values between 0 and 1, in contrast to classical or digital logic, which operates on discrete values of either 1 or 0 (true or false, respectively). Fuzzy logic is widely used in machine control. The term "fuzzy" refers to the fact that the logic involved can deal with concepts that cannot be expressed as the "true" or "false" but rather as "partially true". Although alternative approaches such as genetic algorithms and neural networks can perform just as well as fuzzy logic in many cases, fuzzy logic has the advantage that the solution to the problem can be cast in terms that human operators can understand, such that that their experience can be used in the design of the controller. This makes it easier to mechanize tasks that are already successfully performed by humans. == History and applications == Fuzzy logic was proposed by Lotfi A. Zadeh of the University of California at Berkeley in a 1965 paper. He elaborated on his ideas in a 1973 paper that introduced the concept of "linguistic variables", which in this article equates to a variable defined as a fuzzy set. Other research followed, with the first industrial application, a cement kiln built in Denmark, coming on line in 1976. Fuzzy systems were initially implemented in Japan. Interest in fuzzy systems was sparked by Seiji Yasunobu and Soji Miyamoto of Hitachi, who in 1985 provided simulations that demonstrated the feasibility of fuzzy control systems for the Sendai Subway. Their ideas were adopted, and fuzzy systems were used to control accelerating, braking, and stopping when the Namboku Line opened in 1987. In 1987, Takeshi Yamakawa demonstrated the use of fuzzy control, through a set of simple dedicated fuzzy logic chips, in an "inverted pendulum" experiment. This is a classic control problem, in which a vehicle tries to keep a pole mounted on its top by a hinge upright by moving back and forth. Yamakawa subsequently made the demonstration more sophisticated by mounting a wine glass containing water and even a live mouse to the top of the pendulum: the system maintained stability in both cases. Yamakawa eventually went on to organize his own fuzzy-systems research lab to help exploit his patents in the field. Japanese engineers subsequently developed a wide range of fuzzy systems for both industrial and consumer applications. In 1988 Japan established the Laboratory for International Fuzzy Engineering (LIFE), a cooperative arrangement between 48 companies to pursue fuzzy research. The automotive company Volkswagen was the only foreign corporate member of LIFE, dispatching a researcher for a duration of three years. Japanese consumer goods often incorporate fuzzy systems. Matsushita vacuum cleaners use microcontrollers running fuzzy algorithms to interrogate dust sensors and adjust suction power accordingly. Hitachi washing machines use fuzzy controllers to load-weight, fabric-mix, and dirt sensors and automatically set the wash cycle for the best use of power, water, and detergent. Canon developed an autofocusing camera that uses a charge-coupled device (CCD) to measure the clarity of the image in six regions of its field of view and use the information provided to determine if the image is in focus. It also tracks the rate of change of lens movement during focusing, and controls its speed to prevent overshoot. The camera's fuzzy control system uses 12 inputs: 6 to obtain the current clarity data provided by the CCD and 6 to measure the rate of change of lens movement. The output is the position of the lens. The fuzzy control system uses 13 rules and requires 1.1 kilobytes of memory. An industrial air conditioner designed by Mitsubishi uses 25 heating rules and 25 cooling rules. A temperature sensor provides input, with control outputs fed to an inverter, a compressor valve, and a fan motor. Compared to the previous design, the fuzzy controller heats and cools five times faster, reduces power consumption by 24%, increases temperature stability by a factor of two, and uses fewer sensors. Other applications investigated or implemented include: character and handwriting recognition; optical fuzzy systems; robots, including one for making Japanese flower arrangements; voice-controlled robot helicopters (hovering is a "balancing act" rather similar to the inverted pendulum problem); rehabilitation robotics to provide patient-specific solutions (e.g. to control heart rate and blood pressure ); control of flow of powders in film manufacture; elevator systems; and so on. Work on fuzzy systems is also proceeding in North America and Europe, although on a less extensive scale than in Japan. The US Environmental Protection Agency has investigated fuzzy control for energy-efficient motors, and NASA has studied fuzzy control for automated space docking: simulations show that a fuzzy control system can greatly reduce fuel consumption. Firms such as Boeing, General Motors, Allen-Bradley, Chrysler, Eaton, and Whirlpool have worked on fuzzy logic for use in low-power refrigerators, improved automotive transmissions, and energy-efficient electric motors. In 1995 Maytag introduced an "intelligent" dishwasher based on a fuzzy controller and a "one-stop sensing module" that combines a thermistor, for temperature measurement; a conductivity sensor, to measure detergent level from the ions present in the wash; a turbidity sensor that measures scattered and transmitted light to measure the soiling of the wash; and a magnetostrictive sensor to read spin rate. The system determines the optimum wash cycle for any load to obtain the best results with the least amount of energy, detergent, and water. It even adjusts for dried-on foods by tracking the last time the door was opened, and estimates the number of dishes by the number of times the door was opened. Xiera Technologies Inc. has developed the first auto-tuner for the fuzzy logic controller's knowledge base known as edeX. This technology was tested by Mohawk College and was able to solve non-linear 2x2 and 3x3 multi-input multi-output problems. Research and development is also continuing on fuzzy applications in software, as opposed to firmware, design, including fuzzy expert systems and integration of fuzzy logic with neural-network and so-called adaptive "genetic" software systems, with the ultimate goal of building "self-learning" fuzzy-control systems. These systems can be employed to control complex, nonlinear dynamic plants, for example, human body. == Fuzzy sets == The input variables in a fuzzy control system are in general mapped by sets of membership functions similar to this, known as "fuzzy sets". The process of converting a crisp input value to a fuzzy value is called "fuzzification". The fuzzy logic based approach had been considered by designing two fuzzy systems, one for error heading angle and the other for velocity control. A control system may also have various types of switch, or "ON-OFF", inputs along with its analog inputs, and such switch inputs of course will always have a truth value equal to either 1 or 0, but the scheme can deal with them as simplified fuzzy functions that happen to be either one value or another. Given "mappings" of input variables into membership functions and truth values, the microcontroller then makes decisions for what action to take, based on a set of "rules", each of the form: IF brake temperature IS warm AND speed IS not very fast THEN brake pressure IS slightly decreased. In this example, the two input variables are "brake temperature" and "speed" that have values defined as fuzzy sets. The output variable, "brake pressure" is also defined by a fuzzy set that can have values like "static" or "slightly increased" or "slightly decreased" etc. === Fuzzy control in detail === Fuzzy controllers are very simple conceptually. They consist of an input stage, a processing stage, and an output stage. The input stage maps sensor or other inputs, such as switches, thumbwheels, and so on, to the appropriate membership functions and truth values. The processing stage invokes each appropriate rule and generates a result for each, then combines the results of the rules. Finally, the output stage converts the combined result back into a specific control output value. The most common shape of membership functions is triangular, although trapezoidal and bell curves are also used, but the shape is generally less important than the number of curves and their placement. From three to seven curves are generally appropriate to cover the required range of an input value, or the "universe of discourse" in fuzzy jargon. As discussed earlier, the processing stage is based on a collection of logic rules in the form of IF-THEN statements, where the IF part is called the "antecedent" and the THEN part is called the "consequent". Typical fuzzy

  Read more →
• 

  Apache Pig

  Apache Pig is a high-level platform for creating programs that run on Apache Hadoop. The language for this platform is called Pig Latin. Pig can execute its Hadoop jobs in MapReduce, Apache Tez, or Apache Spark. Pig Latin abstracts the programming from the Java MapReduce idiom into a notation which makes MapReduce programming high level, similar to that of SQL for relational database management systems. Pig Latin can be extended using user-defined functions (UDFs) which the user can write in Java, Python, JavaScript, Ruby or Groovy and then call directly from the language. == History == Apache Pig was originally developed at Yahoo Research around 2006 for researchers to have an ad hoc way of creating and executing MapReduce jobs on very large data sets. In 2007, it was moved into the Apache Software Foundation. === Naming === Regarding the naming of the Pig programming language, the name was chosen arbitrarily and stuck because it was memorable, easy to spell, and for novelty. The story goes that the researchers working on the project initially referred to it simply as 'the language'. Eventually they needed to call it something. Off the top of his head, one researcher suggested Pig, and the name stuck. It is quirky yet memorable and easy to spell. While some have hinted that the name sounds coy or silly, it has provided us with an entertaining nomenclature, such as Pig Latin for the language, Grunt for the shell, and PiggyBank for the CPAN-like shared repository. == Example == Below is an example of a "Word Count" program in Pig Latin: The above program will generate parallel executable tasks which can be distributed across multiple machines in a Hadoop cluster to count the number of words in a dataset such as all the webpages on the internet. == Pig vs SQL == In comparison to SQL, Pig has a nested relational model, uses lazy evaluation, uses extract, transform, load (ETL), is able to store data at any point during a pipeline, declares execution plans, supports pipeline splits, thus allowing workflows to proceed along DAGs instead of strictly sequential pipelines. On the other hand, it has been argued DBMSs are substantially faster than the MapReduce system once the data is loaded, but that loading the data takes considerably longer in the database systems. It has also been argued RDBMSs offer out of the box support for column-storage, working with compressed data, indexes for efficient random data access, and transaction-level fault tolerance. Pig Latin is procedural and fits very naturally in the pipeline paradigm while SQL is instead declarative. In SQL users can specify that data from two tables must be joined, but not what join implementation to use (You can specify the implementation of JOIN in SQL, thus "... for many SQL applications the query writer may not have enough knowledge of the data or enough expertise to specify an appropriate join algorithm."). Pig Latin allows users to specify an implementation or aspects of an implementation to be used in executing a script in several ways. In effect, Pig Latin programming is similar to specifying a query execution plan, making it easier for programmers to explicitly control the flow of their data processing task. SQL is oriented around queries that produce a single result. SQL handles trees naturally, but has no built in mechanism for splitting a data processing stream and applying different operators to each sub-stream. Pig Latin script describes a directed acyclic graph (DAG) rather than a pipeline. Pig Latin's ability to include user code at any point in the pipeline is useful for pipeline development. If SQL is used, data must first be imported into the database, and then the cleansing and transformation process can begin.

  Read more →
• 

  Uncertain inference

  Uncertain inference was first described by C. J. van Rijsbergen as a way to formally define a query and document relationship in Information retrieval. This formalization is a logical implication with an attached measure of uncertainty. == Definitions == Rijsbergen proposes that the measure of uncertainty of a document d to a query q be the probability of its logical implication, i.e.: P ( d → q ) {\displaystyle P(d\to q)} A user's query can be interpreted as a set of assertions about the desired document. It is the system's task to infer, given a particular document, if the query assertions are true. If they are, the document is retrieved. In many cases the contents of documents are not sufficient to assert the queries. A knowledge base of facts and rules is needed, but some of them may be uncertain because there may be a probability associated to using them for inference. Therefore, we can also refer to this as plausible inference. The plausibility of an inference d → q {\displaystyle d\to q} is a function of the plausibility of each query assertion. Rather than retrieving a document that exactly matches the query we should rank the documents based on their plausibility in regards to that query. Since d and q are both generated by users, they are error prone; thus d → q {\displaystyle d\to q} is uncertain. This will affect the plausibility of a given query. By doing this it accomplishes two things: Separate the processes of revising probabilities from the logic Separate the treatment of relevance from the treatment of requests Multimedia documents, like images or videos, have different inference properties for each datatype. They are also different from text document properties. The framework of plausible inference allows us to measure and combine the probabilities coming from these different properties. Uncertain inference generalizes the notions of autoepistemic logic, where truth values are either known or unknown, and when known, they are true or false. == Example == If we have a query of the form: q = A ∧ B ∧ C {\displaystyle q=A\wedge B\wedge C} where A, B and C are query assertions, then for a document D we want the probability: P ( D → ( A ∧ B ∧ C ) ) {\displaystyle P(D\to (A\wedge B\wedge C))} If we transform this into the conditional probability P ( ( A ∧ B ∧ C ) | D ) {\displaystyle P((A\wedge B\wedge C)|D)} and if the query assertions are independent we can calculate the overall probability of the implication as the product of the individual assertions probabilities. == Further work == Croft and Krovetz applied uncertain inference to an information retrieval system for office documents they called OFFICER. In office documents the independence assumption is valid since the query will focus on their individual attributes. Besides analysing the content of documents one can also query about the author, size, topic or collection for example. They devised methods to compare document and query attributes, infer their plausibility and combine it into an overall rating for each document. Besides that uncertainty of document and query contents also had to be addressed. Probabilistic logic networks is a system for performing uncertain inference; crisp true/false truth values are replaced not only by a probability, but also by a confidence level, indicating the certitude of the probability. Markov logic networks allow uncertain inference to be performed; uncertainties are computed using the maximum entropy principle, in analogy to the way that Markov chains describe the uncertainty of finite-state machines.

  Read more →
• 

  AI Action Summit 2025

  The Artificial Intelligence (AI) Action Summit (French: Sommet pour l'action sur l'intelligence artificielle or Sommet pour l'action sur l'IA, SAIA) was held at the Grand Palais in Paris, France, from 10 to 11 February 2025. The summit was co-chaired by French President Emmanuel Macron and Indian Prime Minister Narendra Modi. The 2025 AI Action Summit followed the 2023 AI Safety Summit hosted at Bletchley Park in the UK, and the 2024 AI Seoul Summit in South Korea. This series of AI summits continued with the AI Impact Summit in Delhi, which was hosted by India in February 2026. Whereas the 2023 AI Safety Summit was attended by representatives from 29 governments and executives from only a handful of AI companies, over 1,000 participants from more than 100 countries attended the 2025 Paris AI Summit, representing government leaders, international organisations, the academic and research community, the private sector, and civil society. == Background == The First International AI Safety Report was published on 29 January 2025. Commissioned after the Bletchley Park AI Safety Summit, the report focused on the risks and threats posed by general-purpose AI, and was slated for discussion at the Paris summit as part of the "Trust in AI" pillar. Whereas the first summit was focused on the catastrophic risks of AI and their mitigation, the Paris meeting was recast as an "AI Action Summit" emphasising innovation, practical implementation, and potential economic opportunities of AI, while also exploring a broader range of risks including its environmental impact and disruptions to the labour market. In the weeks leading up to the Paris summit, government leaders had also started to rally around "national champions" in AI, partly in response to Chinese AI startup DeepSeek, which had released a new model rivalling OpenAI o1. On Sunday 9 February, French President Emmanuel Macron posted a compilation of AI-generated deepfake video clips of himself on Instagram to help publicise the start of the 2025 AI Action Summit the following day. While acknowledging the humour of the deepfakes, the real Macron states in the video that using artificial intelligence, "we can do some very big things: change healthcare, energy, life in our society". == Proceedings == === Day 1 === In her opening address, French special envoy Anne Bouverot discussed the environmental impact of AI, acknowledging the technology's "current trajectory is unsustainable". General secretary Christy Hoffman of the UNI Global Union said that "AI-driven productivity gains risk turning the technology into yet another engine of inequality, further straining our democracies". Chinese Vice Premier Zhang Guoqing made a speech expressing China's willingness "to work with other countries to promote development, safeguard security, and share achievements in the field of artificial intelligence". Google CEO Sundar Pichai said in his speech that while the rise of AI brings many risks, "The biggest risk is missing out". He discussed Google's long track record of AI research and said that the company is investing further into "deep research" agents that can autonomously search the Internet and compile a full analysis for users. A new coalition, the Robust Open Online Safety Tools (ROOST) initiative, debuted at the summit. Supported by Google, Discord, OpenAI, and Roblox, and incubated at the Institute of Global Politics at Columbia University, the organisation is developing free, open-source tools to detect and report child sexual abuse material (CSAM). In his speech closing the first day, President Emmanuel Macron emphasized that France has the capability to deliver the power required by AI companies, thanks to its production of nuclear energy. While declaring that Europe was "back in the race" for AI, Macron said that the region was "too slow" for investors, and called on the EU to "simplify regulation" and "resynchronize with the rest of the world". === Day 2 === On 11 February 2025, the French government announced its $400 million endowment of Current AI, a new foundation to support the creation of AI "public goods" including high-quality datasets and open-source tools and infrastructure. Launched by President Macron, Current AI is backed by nine governments – Finland, France, Germany, Chile, India, Kenya, Morocco, Nigeria, Slovenia, and Switzerland – plus various philanthropic organisations such as the Omidyar Group and the McGovern Foundation, and private companies such as Google and Salesforce. Another initiative launched at the summit was the Coalition for Sustainable AI. Led by France, the UN Environment Programme (UNEP), and the International Telecommunication Union (ITU), the coalition has the support of 11 countries, five international organisations, and 37 tech companies including EDF, IBM, Nvidia, and SAP. The Summit of Heads of State and Government took place with a plenary session in the Grand Palais. Prime Minister Narendra Modi of India stressed the need to "democratise technology" and "[ensure] access to all, especially in the Global South". Vice President JD Vance of the United States used his speech to warn against "excessive regulation of the AI" which "could kill a transformative sector just as it's taking off". Vance also warned other leaders against cooperating with "authoritarian regimes" on AI, a comment widely interpreted as a reference to China. == Investments == At the summit, the European Union made several announcements related to planned investments supporting AI development. President Ursula von der Leyen of the European Commission launched InvestAI, a €200 billion initiative, including €20 billion to build four AI gigafactories to train highly complex, very large models. In addition, a coalition of more than 60 European companies launched the EU AI Champions Initiative. Led by venture capital firm General Catalyst, the coalition plans to invest €150 billion in AI-related businesses and infrastructure in Europe over five years. President Emmanuel Macron announced that private investors had pledged to invest nearly €110 billion in the AI sector in France. Financing of between €30 and €50 billion is expected from the United Arab Emirates to build a very large data centre campus, with another €20 billion from the Canadian investment firm Brookfield Corporation. French startup Mistral AI and Helsing, a German-British company, announced their partnership in developing vision-language-action models helping soldiers use AI on the battlefield. == Reactions == The Financial Times editorial board noted that the Paris summit "highlighted a shift in the dynamics towards geopolitical competition", which it characterised as "a new AI arms race" between the US and China, with Europe "trying to carve out its role". Fortune.com AI editor Jeremy Kahn described the 2025 Paris Summit as an "AI festival, complete with glitzy corporate side events and even a late night dance party", contrasting it with the "decidedly sober" mood of the inaugural AI Safety Summit at Bletchley Park. Many experts of the AI Safety Community expressed disappointment that the Paris Summit did not do enough to address AI risks, with Anthropic CEO Dario Amodei calling it a "missed opportunity". Others voicing similar concerns included David Leslie of the Alan Turing Institute and Max Tegmark of the Future of Life Institute. Reporting from Paris, technology columnist Kevin Roose of The New York Times wrote, "The biggest surprise of the Paris summit, for me, has been that policymakers can't seem to grasp how soon powerful AI systems could arrive, or how disruptive they could be." == Statement on inclusive and sustainable AI == At the summit, 58 countries, including France, China, and India, signed a joint declaration, the Statement on Inclusive and Sustainable Artificial Intelligence for People and the Planet. The statement outlines general principles such as accessibility and overcoming the digital divide; developing AI that is open, transparent, ethical, safe, and trustworthy; avoiding market concentration of AI development to encourage innovation; positive outcomes for labour markets; making AI sustainable; and promoting international cooperation and governance. The US and UK refused to sign the declaration on inclusive and sustainable AI. The UK government said in a brief statement that the international agreement did not go far enough in defining global governance of AI and addressing concerns about its impact on national security. === Signatories === The list of signatory countries to the statement for inclusive and sustainable AI in alphabetical order: Additional signatories included the following international bodies and research institutes: ALAI (Latin American Association on Internet) African Union (AU) Commission BEUC The European Consumer Organisation Center for Democracy and Technology Council of Europe European Commission (and the 27 member states) Hugging Face INRIA Institute of Advanced Study OEC

  Read more →