Dynamic Bayesian network

A dynamic Bayesian network (DBN) is a Bayesian network (BN) which relates variables to each other over adjacent time steps. == History == A dynamic Bayesian network (DBN) is often called a "two-timeslice" BN (2TBN) because it says that at any point in time T, the value of a variable can be calculated from the internal regressors and the immediate prior value (time T-1). DBNs were developed by Paul Dagum in the early 1990s at Stanford University's Section on Medical Informatics. Dagum developed DBNs to unify and extend traditional linear state-space models such as Kalman filters, linear and normal forecasting models such as ARMA and simple dependency models such as hidden Markov models into a general probabilistic representation and inference mechanism for arbitrary nonlinear and non-normal time-dependent domains. Today, DBNs are common in robotics, and have shown potential for a wide range of data mining applications. For example, they have been used in speech recognition, digital forensics, protein sequencing, and bioinformatics. DBN is a generalization of hidden Markov models and Kalman filters. DBNs are conceptually related to probabilistic Boolean networks and can, similarly, be used to model dynamical systems at steady-state.

ParkMobile

ParkMobile is a mobile and web app providing parking payments in North America. Headquartered in Atlanta, Georgia, users can pay for on-street and off-street parking via app on their smartphone, web browser, or through calling a phone number. ParkMobile also offers parking reservations at stadiums or venues for concerts and sporting events, and in metro area garages. == History == ParkMobile was founded in the United States in 2008 by Albert Bogaard after originally starting in the Netherlands. The initial product served only zone (on-demand) parkers and payment for the parking spot was made via a phone call through an IVR system. In 2009, the ParkMobile app was released and the product launched in its first city, Grand Rapids, Michigan. Parking payments have since been accepted through a user's account by connecting a credit card. ParkMobile deployed in Washington, D.C., in 2011. As of 2023, ParkMobile now has over 50 million users. Parking reservations were introduced in 2017, allowing users to reserve parking in advance. In 2018, the company recapitalized with BMW as the shareholder. ParkMobile was then acquired by a joint venture with BMW and Daimler. Under this joint venture, ParkMobile parking payment functionality was available and integrated with BMW's navigation system in many of its 2018 models. EasyPark Group, the Swedish-based parking solutions company, acquired ParkMobile in 2021 and is the current owner rebranded as Arrive. In 2022, ParkMobile launched in the City of Boston with a city-wide parking app, ParkBoston, powered by ParkMobile. == Operations == === Products === ParkMobile's product offerings include zone (on-demand) parking payments, parking reservations, and a self-service reporting engine. Zone parking is the company's most widely used service. Users can use the app on their smartphone to pay parking fees. In 2017, ParkMobile began offering parking reservations. The service is provided in addition to on-demand parking options at stadiums and venues, as well as metro area parking garages. After launching the reservations feature, ParkMobile became the first mobile parking app provider in North America to have a consolidated app with both on-demand and reservations parking in one. ParkMobile 360, the company's self-service management and reporting platform for operators, launched in 2018. It is a web-based application for parking operators to manage parking inventory, adjust rates, create special parking events, and track analytics. In 2020, ParkMobile began offering an option to pay for parking with Google through integrating the ParkMobile experience with Google Maps In 2021, ParkMobile launched its web application, allowing users to complete their parking transactions directly from the mobile website without having to download the app or have an account. ParkMobile integrates with parking gate equipment so customers can use their app to pay for parking and scan to enter and exit the garage. === Locations === ParkMobile has over 50 million users across the United States, Canada, and Puerto Rico. The app is available in over 550 cities in the U.S. and over 150 colleges and universities. == Controversies == === Predatory towing and excessive ticketing === Since all paid parking sessions from a single supplier are able to be viewed together, the ease of viewing and enforcing parking violations has caused controversy. Parking Enforcement Services in Birmingham, Alabama, has been the subject complaints by users of the ParkMobile app who had paid for a parking session and still had their vehicle towed. Customers often use old or expired license plates and forget to update to the correct number, or mistype when entering their information into the ParkMobile app. The complaints are that the towing companies offer no lenience for these mistakes. They return to their car as the session expires, and find their car has been towed. Additionally, other municipality across the country have received complaints about excessive parking ticket issuing when inputting their information incorrectly in the ParkMobile app. In Stone Harbor, New Jersey, parking ticket violations increased by over 1,600% from the previous year since launching with the ParkMobile app. Police officers refute complaints of being "too strict" on writing tickets by admitting the ParkMobile system allows officers to "more seamlessly enforce" the city's parking laws. === Data security breach === In March 2021, ParkMobile suffered a cybersecurity incident "linked to a vulnerability in a third-party software," potentially exposing users' email addresses, phone numbers, and license plate numbers. ParkMobile responded by launching an investigation and notifying law enforcement authorities and affected municipalities. The investigation concluded "no sensitive data or Payment Card Information was affected" but ParkMobile confirmed that basic account information, such as license plate numbers and possibly email addresses or phone numbers, was accessed.

Characteristic samples

Characteristic samples is a concept in the field of grammatical inference, related to passive learning. In passive learning, an inference algorithm I {\displaystyle I} is given a set of pairs of strings and labels S {\displaystyle S} , and returns a representation R {\displaystyle R} that is consistent with S {\displaystyle S} . Characteristic samples consider the scenario when the goal is not only finding a representation consistent with S {\displaystyle S} , but finding a representation that recognizes a specific target language. A characteristic sample of language L {\displaystyle L} is a set of pairs of the form ( s , l ( s ) ) {\displaystyle (s,l(s))} where: l ( s ) = 1 {\displaystyle l(s)=1} if and only if s ∈ L {\displaystyle s\in L} l ( s ) = − 1 {\displaystyle l(s)=-1} if and only if s ∉ L {\displaystyle s\notin L} Given the characteristic sample S {\displaystyle S} , I {\displaystyle I} 's output on it is a representation R {\displaystyle R} , e.g. an automaton, that recognizes L {\displaystyle L} . == Formal Definition == === The Learning Paradigm associated with Characteristic Samples === There are three entities in the learning paradigm connected to characteristic samples, the adversary, the teacher and the inference algorithm. Given a class of languages C {\displaystyle \mathbb {C} } and a class of representations for the languages R {\displaystyle \mathbb {R} } , the paradigm goes as follows: The adversary A {\displaystyle A} selects a language L ∈ C {\displaystyle L\in \mathbb {C} } and reports it to the teacher The teacher T {\displaystyle T} then computes a set of strings and label them correctly according to L {\displaystyle L} , trying to make sure that the inference algorithm will compute L {\displaystyle L} The adversary can add correctly labeled words to the set in order to confuse the inference algorithm The inference algorithm I {\displaystyle I} gets the sample and computes a representation R ∈ R {\displaystyle R\in \mathbb {R} } consistent with the sample. The goal is that when the inference algorithm receives a characteristic sample for a language L {\displaystyle L} , or a sample that subsumes a characteristic sample for L {\displaystyle L} , it will return a representation that recognizes exactly the language L {\displaystyle L} . === Sample === Sample S {\displaystyle S} is a set of pairs of the form ( s , l ( s ) ) {\displaystyle (s,l(s))} such that l ( s ) ∈ { − 1 , 1 } {\displaystyle l(s)\in \{-1,1\}} ==== Sample consistent with a language ==== We say that a sample S {\displaystyle S} is consistent with language L {\displaystyle L} if for every pair ( s , l ( s ) ) {\displaystyle (s,l(s))} in S {\displaystyle S} : l ( s ) = 1 if and only if s ∈ L {\displaystyle l(s)=1{\text{ if and only if }}s\in L} l ( s ) = − 1 if and only if s ∉ L {\displaystyle l(s)=-1{\text{ if and only if }}s\notin L} === Characteristic sample === Given an inference algorithm I {\displaystyle I} and a language L {\displaystyle L} , a sample S {\displaystyle S} that is consistent with L {\displaystyle L} is called a characteristic sample of L {\displaystyle L} for I {\displaystyle I} if: I {\displaystyle I} 's output on S {\displaystyle S} is a representation R {\displaystyle R} that recognizes L {\displaystyle L} . For every sample D {\displaystyle D} that is consistent with L {\displaystyle L} and also fulfils S ⊆ D {\displaystyle S\subseteq D} , I {\displaystyle I} 's output on D {\displaystyle D} is a representation R {\displaystyle R} that recognizes L {\displaystyle L} . A Class of languages C {\displaystyle \mathbb {C} } is said to have charistaristic samples if every L ∈ C {\displaystyle L\in \mathbb {C} } has a characteristic sample. == Related Theorems == === Theorem === If equivalence is undecidable for a class C {\textstyle \mathbb {C} } over Σ {\textstyle \Sigma } of cardinality bigger than 1, then C {\textstyle \mathbb {C} } doesn't have characteristic samples. ==== Proof ==== Given a class of representations C {\textstyle \mathbb {C} } such that equivalence is undecidable, for every polynomial p ( x ) {\displaystyle p(x)} and every n ∈ N {\displaystyle n\in \mathbb {N} } , there exist two representations r 1 {\displaystyle r_{1}} and r 2 {\displaystyle r_{2}} of sizes bounded by n {\displaystyle n} , that recognize different languages but are inseparable by any string of size bounded by p ( n ) {\displaystyle p(n)} . Assuming this is not the case, we can decide if r 1 {\displaystyle r_{1}} and r 2 {\displaystyle r_{2}} are equivalent by simulating their run on all strings of size smaller than p ( n ) {\displaystyle p(n)} , contradicting the assumption that equivalence is undecidable. === Theorem === If S 1 {\displaystyle S_{1}} is a characteristic sample for L 1 {\displaystyle L_{1}} and is also consistent with L 2 {\displaystyle L_{2}} , then every characteristic sample of L 2 {\displaystyle L_{2}} , is inconsistent with L 1 {\displaystyle L_{1}} . ==== Proof ==== Given a class C {\textstyle \mathbb {C} } that has characteristic samples, let R 1 {\displaystyle R_{1}} and R 2 {\displaystyle R_{2}} be representations that recognize L 1 {\displaystyle L_{1}} and L 2 {\displaystyle L_{2}} respectively. Under the assumption that there is a characteristic sample for L 1 {\displaystyle L_{1}} , S 1 {\displaystyle S_{1}} that is also consistent with L 2 {\displaystyle L_{2}} , we'll assume falsely that there exist a characteristic sample for L 2 {\displaystyle L_{2}} , S 2 {\displaystyle S_{2}} that is consistent with L 1 {\displaystyle L_{1}} . By the definition of characteristic sample, the inference algorithm I {\displaystyle I} must return a representation which recognizes the language if given a sample that subsumes the characteristic sample itself. But for the sample S 1 ∪ S 2 {\displaystyle S_{1}\cup S_{2}} , the answer of the inferring algorithm needs to recognize both L 1 {\displaystyle L_{1}} and L 2 {\displaystyle L_{2}} , in contradiction. === Theorem === If a class is polynomially learnable by example based queries, it is learnable with characteristic samples. == Polynomialy characterizable classes == === Regular languages === The proof that DFA's are learnable using characteristic samples, relies on the fact that every regular language has a finite number of equivalence classes with respect to the right congruence relation, ∼ L {\displaystyle \sim _{L}} (where x ∼ L y {\displaystyle x\sim _{L}y} for x , y ∈ Σ ∗ {\displaystyle x,y\in \Sigma ^{}} if and only if ∀ z ∈ Σ ∗ : x z ∈ L ↔ y z ∈ L {\displaystyle \forall z\in \Sigma ^{}:xz\in L\leftrightarrow yz\in L} ). Note that if x {\displaystyle x} , y {\displaystyle y} are not congruent with respect to ∼ L {\displaystyle \sim _{L}} , there exists a string z {\displaystyle z} such that x z ∈ L {\displaystyle xz\in L} but y z ∉ L {\displaystyle yz\notin L} or vice versa, this string is called a separating suffix. ==== Constructing a characteristic sample ==== The construction of a characteristic sample for a language L {\displaystyle L} by the teacher goes as follows. Firstly, by running a depth first search on a deterministic automaton A {\displaystyle A} recognizing L {\displaystyle L} , starting from its initial state, we get a suffix closed set of words, W {\displaystyle W} , ordered in shortlex order. From the fact above, we know that for every two states in the automaton, there exists a separating suffix that separates between every two strings that the run of A {\displaystyle A} on them ends in the respective states. We refer to the set of separating suffixes as S {\displaystyle S} . The labeled set (sample) of words the teacher gives the adversary is { ( w , l ( w ) ) | w ∈ W ⋅ S ∪ W ⋅ Σ ⋅ S } {\displaystyle \{(w,l(w))|w\in W\cdot S\cup W\cdot \Sigma \cdot S\}} where l ( w ) {\displaystyle l(w)} is the correct label of w {\displaystyle w} (whether it is in L {\displaystyle L} or not). We may assume that ϵ ∈ S {\displaystyle \epsilon \in S} . ==== Constructing a deterministic automata ==== Given the sample from the adversary W {\displaystyle W} , the construction of the automaton by the inference algorithm I {\displaystyle I} starts with defining P = prefix ( W ) {\displaystyle P={\text{prefix}}(W)} and S = suffix ( W ) {\displaystyle S={\text{suffix}}(W)} , which are the set of prefixes and suffixes of W {\displaystyle W} respectively. Now the algorithm constructs a matrix M {\displaystyle M} where the elements of P {\displaystyle P} function as the rows, ordered by the shortlex order, and the elements of S {\displaystyle S} function as the columns, ordered by the shortlex order. Next, the cells in the matrix are filled in the following manner for prefix p i {\displaystyle p_{i}} and suffix s j {\displaystyle s_{j}} : If p i s j ∈ W → M i j = l ( p i s j ) {\displaystyle p_{i}s_{j}\in W\rightarrow M_{ij}=l(p_{i}s_{j})} else, M i j = 0 {\displaystyle M_{ij}=0} Now, we say row i {\displaystyle i} and t {\displaystyle t} are distinguishable if there exi

Soft independent modelling of class analogies

Soft independent modelling by class analogy (SIMCA) is a statistical method for supervised classification of data. The method requires a training data set consisting of samples (or objects) with a set of attributes and their class membership. The term soft refers to the fact the classifier can identify samples as belonging to multiple classes and not necessarily producing a classification of samples into non-overlapping classes. == Method == In order to build the classification models, the samples belonging to each class need to be analysed using principal component analysis (PCA); only the significant components are retained. For a given class, the resulting model then describes either a line (for one Principal Component or PC), plane (for two PCs) or hyper-plane (for more than two PCs). For each modelled class, the mean orthogonal distance of training data samples from the line, plane, or hyper-plane (calculated as the residual standard deviation) is used to determine a critical distance for classification. This critical distance is based on the F-distribution and is usually calculated using 95% or 99% confidence intervals. New observations are projected into each PC model and the residual distances calculated. An observation is assigned to the model class when its residual distance from the model is below the statistical limit for the class. The observation may be found to belong to multiple classes and a measure of goodness of the model can be found from the number of cases where the observations are classified into multiple classes. The classification efficiency is usually indicated by Receiver operating characteristics. In the original SIMCA method, the ends of the hyper-plane of each class are closed off by setting statistical control limits along the retained principal components axes (i.e., score value between plus and minus 0.5 times score standard deviation). More recent adaptations of the SIMCA method close off the hyper-plane by construction of ellipsoids (e.g. Hotelling's T2 or Mahalanobis distance). With such modified SIMCA methods, classification of an object requires both that its orthogonal distance from the model and its projection within the model (i.e. score value within the region defined by the ellipsoid) are not significant. == Application == SIMCA as a method of classification has gained widespread use especially in applied statistical fields such as chemometrics and spectroscopic data analysis.

Constructing skill trees

Constructing skill trees (CST) is a hierarchical reinforcement learning algorithm which can build skill trees from a set of sample solution trajectories obtained from demonstration. CST uses an incremental MAP (maximum a posteriori) change point detection algorithm to segment each demonstration trajectory into skills and integrate the results into a skill tree. CST was introduced by George Konidaris, Scott Kuindersma, Andrew Barto and Roderic Grupen in 2010. == Algorithm == CST consists of mainly three parts;change point detection, alignment and merging. The main focus of CST is online change-point detection. The change-point detection algorithm is used to segment data into skills and uses the sum of discounted reward R t {\displaystyle R_{t}} as the target regression variable. Each skill is assigned an appropriate abstraction. A particle filter is used to control the computational complexity of CST. The change point detection algorithm is implemented as follows. The data for times t ∈ T {\displaystyle t\in T} and models Q with prior p ( q ∈ Q ) {\displaystyle p(q\in Q)} are given. The algorithm is assumed to be able to fit a segment from time j + 1 {\displaystyle j+1} to t using model q with the fit probability P ( j , t , q ) {\displaystyle P(j,t,q)_{}^{}} . A linear regression model with Gaussian noise is used to compute P ( j , t , q ) {\displaystyle P(j,t,q)} . The Gaussian noise prior has mean zero, and variance which follows I n v e r s e G a m m a ( v 2 , u 2 ) {\displaystyle \mathrm {InverseGamma} \left({\frac {v}{2}},{\frac {u}{2}}\right)} . The prior for each weight follows N o r m a l ( 0 , σ 2 δ ) {\displaystyle \mathrm {Normal} (0,\sigma ^{2}\delta )} . The fit probability P ( j , t , q ) {\displaystyle P(j,t,q)} is computed by the following equation. P ( j , t , q ) = π − n 2 δ m | ( A + D ) − 1 | 1 2 u v 2 ( y + u ) u + v 2 Γ ( n + v 2 ) Γ ( v 2 ) {\displaystyle P(j,t,q)={\frac {\pi ^{-{\frac {n}{2}}}}{\delta ^{m}}}\left|(A+D)^{-1}\right|^{\frac {1}{2}}{\frac {u^{\frac {v}{2}}}{(y+u)^{\frac {u+v}{2}}}}{\frac {\Gamma ({\frac {n+v}{2}})}{\Gamma ({\frac {v}{2}})}}} Then, CST compute the probability of the changepoint at time j with model q, P t ( j , q ) {\displaystyle P_{t}(j,q)} and P j MAP {\displaystyle P_{j}^{\text{MAP}}} using a Viterbi algorithm. P t ( j , q ) = ( 1 − G ( t − j − 1 ) ) P ( j , t , q ) p ( q ) P j MAP {\displaystyle P_{t}(j,q)=(1-G(t-j-1))P(j,t,q)p(q)P_{j}^{\text{MAP}}} P j MAP = max i , q P j ( i , q ) g ( j − i ) 1 − G ( j − i − 1 ) , ∀ j < t {\displaystyle P_{j}^{\text{MAP}}=\max _{i,q}{\frac {P_{j}(i,q)g(j-i)}{1-G(j-i-1)}},\forall j

SEMAT

SEMAT (Software Engineering Method and Theory) is an initiative to reshape software engineering such that software engineering qualifies as a rigorous discipline. The initiative was launched in December 2009 by Ivar Jacobson, Bertrand Meyer, and Richard Soley with a call for action statement and a vision statement. The initiative was envisioned as a multi-year effort for bridging the gap between the developer community and the academic community and for creating a community giving value to the whole software community. The work is now structured in four different but strongly related areas: Practice, Education, Theory, and Community. The Practice area primarily addresses practices. The Education area is concerned with all issues related to training for both the developers and the academics including students. The Theory area is primarily addressing the search for a General Theory in Software Engineering. Finally, the Community area works with setting up legal entities, creating websites and community growth. It was expected that the Practice area, the Education area and the Theory area would at some point in time integrate in a way of value to all of them: the Practice area would be a "customer" of the Theory area, and direct the research to useful results for the developer community. The Theory area would give a solid and practical platform for the Practice area. And, the Education area would communicate the results in proper ways. == Practice area == The first step was here to develop a common ground or a kernel including the essence of software engineering – things we always have, always do, always produce when developing software. The second step was envisioned to add value on top of this kernel in the form of a library of practices to be composed to become specific methods, specific for all kinds of reasons such as the preferences of the team using it, kind of software being built, etc. The first step is as of this writing just about to be concluded. The results are a kernel including universal elements for software development – called the Essence Kernel, and a language – called the Essence Language - to describe these elements (and elements built on top of the kernel (practices, methods, and more). Essence, including both the kernel and language, has been published as an OMG standard in beta status in July 2013 and is expected to become a formally adopted standard in early 2014. The second step has just started, and the Practice area will be divided into a number of separate but interconnected tracks: the practice (library track), the tool track are so far identified and work has started or is about to get started. The practice track is currently working on a Users Guide. == Education area == The area focuses on leveraging the work of SEMAT in software engineering education, both within academia and industry. It promotes global education based on a common ground called Essence. The area's target groups are instructors such as university professors and industrial coaches as well as their students and learning practitioners. The goal of the area is to create educational courses and course materials that are internationally viable, identify pedagogical approaches that are appropriate and effective for specific target groups and disseminate experience and lessons learned. The area includes members from a number of universities and institutes worldwide. Most members have already been involved in leveraging aspects of SEMAT in the context of their software engineering courses. They are gathering their resources and starting a common venture towards defining a new generation of SEMAT-powered software engineering curricula. As of 2018, some studies of utilizing Essence in educational settings exist. One example of the use of Essence in university education was a software engineering course carried out in Norwegian University of Science and Technology. A study was conducted by introducing Essence into a project-based software engineering course, with the aim of understanding what difficulties the students faced in using Essence, and whether they considered it to have been useful. The results indicated that Essence could also be useful for novice software engineers by (1) encouraging them to look up and study new practices and methods in order to create their own, (2) encouraging them to adjust their way-of-working reflectively and in a situation-specific manner, (3) helping them structure their way of working. The findings of another study introducing students to Essence through a digital game supported these findings: the students felt that Essence will be useful to them in future, real-world projects, and that they wish to utilize it in them. == Theory area == An important part of SEMAT is that a general theory of software engineering is planned to emerge with significant benefits. A series of workshops held under the title SEMAT Workshop on a General Theory of Software Engineering (GTSE) are a key component in awareness building around general theories. In addition to community awareness building, SEMAT also aims to contribute with a specific general theory of software engineering. This theory should be solidly based on the SEMAT Essence language and kernel, and should support software engineering practitioners' goal-oriented decision making. As argued elsewhere, such support is predicated on the predictive capabilities of the theory. Thus, the SEMAT Essence should be augmented to allow the prediction of critical software engineering phenomena. The GTSE workshop series assists in the development of the SEMAT general software engineering theory by engaging a larger community in the search for, development of, and evaluation of promising theories, which may be used as a base for the SEMAT theory. == Organizational structure == === Main organization === SEMAT is chaired by Sumeet S. Malhotra of Tata Consultancy Services. The CEO of the organization is Ste Nadin of Fujitsu. The Executive Management Committee of SEMAT are Ivar Jacobson, Ste Nadin, Sumeet S. Malhotra, Paul E. McMahon, Michael Goedicke and Cecile Peraire. === Japan Chapter === Japan Chapter was established in April 2013, and it has more than 250 members as of November 2013. Member activities include carrying out seminars about SEMAT, considering utilization of SEMAT Essence for integrating different requirements engineering techniques and body of knowledges (BoKs), and translating articles into Japanese. === Korea Chapter === The chapter was inaugurated with about 50 members in October 2013. Member activities include: 2e Consulting started rewriting their IT service engagement methods using the Essence kernel, and uEngine Solutions started developing a tool to orchestrate Essence-kernel based practices into a project method. Korean government supported KAIST to conduct research in Essence. === Latin American Chapter === Semat Latin American Chapter was created in August 2011 in Medellin (Colombia) by Ivar Jacobson during the Latin American Software Engineering Symposium. This Chapter has 9 Executive Committee members from Colombia, Venezuela, Peru, Brazil, Argentina, Chile, and Mexico, chaired by Dr. Carlos Zapata from Colombia. More than 80 people signed the initial declaration of the Chapter and nowadays the Chapter members are in charge of disseminating the Semat ideas in all Latin America. Chapter members have participated in various Latin American conferences, including the Latin American Conference on Informatics (CLEI), the Ibero American Software Engineering and Knowledge Engineering Journeys (JIISIC), the Colombian Computing Conference (CCC), and the Chilean Computing Meeting (ECC). The Chapter contributed in the submission sent in response to the OMG call for proposals and currently studies didactic strategies for teaching the Semat kernel by games, theoretical studies about some kernel elements, and practical representations of several software development and quality methods by using the Semat kernel. Some of the members also translated the Essence book and some other Semat materials and papers into Spanish. === Russia Chapter === Russian Chapter has about 20 members. A few universities have incorporated SEMAT in their training courses , including Moscow State University, Moscow Institute of Physics and Technology, Higher School of Economics, Moscow State University of Economics, Statistics, and Informatics. The chapter and some commercial companies are carrying out seminars about SEMAT. INCOSE Russian Chapter is working on an extension of SEMAT to systems engineering. EC-leasing is working on an extension of the Kernel for Software Life Cycle. Russian Chapter attended in two conferences: Actual Problems of System and Software Engineering and SECR with SEMAT section and articles. Translation of the Essence book into Russian is in progress. == Practical Applications of SEMAT == Ideas developed by the SEMAT community have been applied by both industry and ac

Multimodal learning

Multimodal learning is a type of deep learning that integrates and processes multiple types of data, referred to as modalities, such as text, audio, images, or video. This integration allows for a more holistic understanding of complex data, improving model performance in tasks like visual question answering, cross-modal retrieval, text-to-image generation, aesthetic ranking, and image captioning. Multimodal learning was proposed in 2011 at the beginning of the deep learning period. Large multimodal models, such as Google Gemini and GPT-4o, have become increasingly popular since 2023, enabling increased versatility and a broader understanding of real-world phenomena. == Motivation == Data usually comes with different modalities which carry different information. For example, it is very common to caption an image to convey the information not presented in the image itself. Similarly, sometimes it is more straightforward to use an image to describe information which may not be obvious from text. As a result, if different words appear in similar images, then these words likely describe the same thing. Conversely, if a word is used to describe seemingly dissimilar images, then these images may represent the same object. Thus, in cases dealing with multi-modal data, it is important to use a model which is able to jointly represent the information such that the model can capture the combined information from different modalities. == Multimodal transformers == Models such as CLIP (Contrastive Language–Image Pretraining) learn joint representations of images and text by optimizing contrastive objectives, allowing the model to match images with their corresponding textual descriptions. == Multimodal deep Boltzmann machines == A Boltzmann machine is a type of stochastic neural network invented by Geoffrey Hinton and Terry Sejnowski in 1985. Boltzmann machines can be seen as the stochastic, generative counterpart of Hopfield nets. They are named after the Boltzmann distribution in statistical mechanics. The units in Boltzmann machines are divided into two groups: visible units and hidden units. Each unit is like a neuron with a binary output that represents whether it is activated or not. General Boltzmann machines allow connection between any units. However, learning is impractical using general Boltzmann Machines because the computational time is exponential to the size of the machine. A more efficient architecture is called restricted Boltzmann machine where connection is only allowed between hidden unit and visible unit, which is described in the next section. Multimodal deep Boltzmann machines can process and learn from different types of information, such as images and text, simultaneously. This can notably be done by having a separate deep Boltzmann machine for each modality, for example one for images and one for text, joined at an additional top hidden layer. == Applications == Multimodal machine learning has numerous applications across various domains: Cross-modal retrieval: cross-modal retrieval allows users to search for data across different modalities (e.g., retrieving images based on text descriptions), improving multimedia search engines and content recommendation systems. Classification and missing data retrieval: multimodal Deep Boltzmann Machines outperform traditional models like support vector machines and latent Dirichlet allocation in classification tasks and can predict missing data in multimodal datasets, such as images and text. Healthcare diagnostics: multimodal models integrate medical imaging, genomic data, and patient records to improve diagnostic accuracy and early disease detection, especially in cancer screening. Content generation: models like DALL·E generate images from textual descriptions, benefiting creative industries, while cross-modal retrieval enables dynamic multimedia searches. Robotics and human-computer interaction: multimodal learning improves interaction in robotics and AI by integrating sensory inputs like speech, vision, and touch, aiding autonomous systems and human-computer interaction. Emotion recognition: combining visual, audio, and text data, multimodal systems enhance sentiment analysis and emotion recognition, applied in customer service, social media, and marketing.