Alexey Yakovlevich Chervonenkis (Russian: Алексей Яковлевич Червоненкис; 7 September 1938 – 22 September 2014) was a Soviet and Russian mathematician. Along with Vladimir Vapnik, he was one of the main developers of the Vapnik–Chervonenkis theory, also known as the "fundamental theory of learning", an important part of computational learning theory. Chervonenkis held joint appointments with the Russian Academy of Sciences and Royal Holloway, University of London. Alexey Chervonenkis got lost in Losiny Ostrov National Park on 22 September 2014, and later during a search operation was found dead near Mytishchi, a suburb of Moscow. He had died of hypothermia.
Stixel
In computer vision, a stixel (portmanteau of "stick" and "pixel") is a superpixel representation of depth information in an image, in the form of a vertical stick that approximates the closest obstacles within a certain vertical slice of the scene. Introduced in 2009, stixels have applications in robotic navigation and advanced driver-assistance systems, where they can be used to define a representation of robotic environments and traffic scenes with a medium level of abstraction. == Definition == One of the problems of scene understanding in computer vision is to determine horizontal freespace around the camera, where the agent can move, and the vertical obstacles delimiting it. An image can be paired with depth information (produced e.g. from stereo disparity, lidar, or monocular depth estimation), allowing a dense tridimensional reconstruction of the observed scene. One drawback of dense reconstruction is the large amount of data involved, since each pixel in the image is mapped to an element of a point cloud. Vision problems characterised by planar freespace delimited by mostly vertical obstacles, such as traffic scenes or robotic navigation, can benefit from a condensed representation that allows to save memory and processing time. Stixels are thin vertical rectangles representing a slice of a vertical surface belonging to the closest obstacle in the observed scene. They allow to dramatically reduce the amount of information needed to represent a scene in such problems. A stixel is characterised by three parameters: vertical coordinate of the bottom, height of the stick, and depth. Stixels have fixed width, with each stixel spanning over a certain number of image columns, allowing downsampling of the horizontal image resolution. In the original formulation, each column of the image would contain at most one stixel, and later extensions were developed to allow multiple stixels on each column, allowing to represent multiple objects at different distances. == Stixel estimation == The input to stixel estimation is a dense depth map, that can be computed from stereo disparity or other means. The original approach computes an occupancy grid that can be segmented to estimate the freespace, with dynamic programming providing an efficient method to find an optimal segmentation. Alternative approaches can be used instead of occupancy grid mapping, such as manifold-based methods. The freespace boundary provides the base points of the obstacles at closest longitudinal distance, however multiple objects at different distances might appear in each column of the image. To fully define the obstacles, their height should be estimated, and this is accomplished by segmenting the depth of the object from the depth of the background. A membership function over the pixels can be defined based on the depth value, where the membership represents the confidence of a pixel belonging to the closest vertical obstacle or to the background, and a cut separating the obstacles from the background can again be computed effectively with dynamic programming. Once both the freespace and the obstacle height are known, the stixels can be estimated by fusing the information over the columns spanned by each stixel, and finally a refined depth of the stixel can be estimated via model fitting over the depth of the pixels covered by the stixel, possibly paired with confidence information (e.g. disparity confidence produced by methods such as semi-global matching).
AI Photo Editors: Free vs Paid (2026)
Trying to pick the best AI photo editor? An AI photo editor is software that uses machine learning to help you get more done — it scales effortlessly from a single task to thousands. The best picks balance beginner-friendly simplicity with the depth power users need, and they ship updates often. Whether you are a beginner or a pro, the right AI photo editor slots into your workflow and pays for itself fast. Read on for hands-on impressions, pricing tiers, and the standout features that matter.
Vasant Honavar
Vasant G. Honavar is an Indian-American computer scientist, and artificial intelligence, machine learning, big data, data science, causal inference, knowledge representation, bioinformatics and health informatics researcher and professor. == Early life and education == Vasant Honavar was born at Pune, India to Bhavani G. and Gajanan N. Honavar. He received his early education at the Vidya Vardhaka Sangha High School and M.E.S. College in Bangalore, India. He received a B.E. in Electronics & Communications Engineering from the B.M.S. College of Engineering in Bangalore, India in 1982, when it was affiliated with Bangalore University, an M.S. in electrical and computer engineering in 1984 from Drexel University, and an M.S. in computer science in 1989, and a Ph.D. in 1990, respectively, from the University of Wisconsin–Madison, where he studied Artificial Intelligence and worked with Leonard Uhr. == Career == Honavar is on the faculty of Informatics and Intelligent Systems Department in the Penn State College of Information Sciences and Technology at Pennsylvania State University where he currently holds the Dorothy Foehr Huck and J. Lloyd Huck Chair in Biomedical Data Sciences and Artificial Intelligence and previously held the Edward Frymoyer Endowed Chair in Information Sciences and Technology. He serves on the faculties of the graduate programs in Computer Science, Informatics, Bioinformatics and Genomics, Neuroscience, Operations Research, Public Health Sciences, and of undergraduate programs in Data Science and Artificial Intelligence methods and applications. Honavar serves as the director of the Artificial Intelligence Research Laboratory, Director of Strategic Initiatives for the Institute for Computational and Data Sciences and the director of the Center for Artificial Intelligence Foundations and Scientific Applications at Pennsylvania State University. Honavar served on the Leadership Team of the Northeast Big Data Innovation Hub. Honavar served on the Computing Research Association's Computing Community Consortium Council during 2014-2017, where he chaired the task force on Convergence of Data and Computing, and was a member of the task force on Artificial Intelligence. Honavar was the first Sudha Murty Distinguished Visiting Chair of Neurocomputing and Data Science by the Indian Institute of Science, Bangalore, India. Honavar was named a Distinguished Member of the Association for Computing Machinery for "outstanding scientific contributions to computing"; and elected a Fellow of the American Association for the Advancement of Science for his "distinguished research contributions and leadership in data science". As a Program Director in the Information Integration and Informatics program in the Information and Intelligent Systems Division of the Computer and Information Science and Engineering Directorate of the US National Science Foundation during 2010-13, Honavar led the Big Data Program. Honavar was a professor of computer science at Iowa State University where he led the Artificial Intelligence Research Laboratory which he founded in 1990 and was instrumental in establishing an interdepartmental graduate program in Bioinformatics and Computational Biology (and served as its Chair during 2003–2005). Honavar has held visiting professorships at Carnegie Mellon University, the University of Wisconsin–Madison, and at the Indian Institute of Science. == Research == Honavar's research has contributed to advances in artificial intelligence, machine learning, causal inference, knowledge representation, neural networks, semantic web, big data analytics, and bioinformatics and computational biology. He was a program chair of the Association for the Advancement of Artificial Intelligence(AAAI)'s 36th Conference on Artificial Intelligence. He has published over 300 research articles, including many highly cited ones, as well as several books on these topics. His recent work has focused on federated machine learning algorithms for constructing predictive models from distributed data and linked open data, learning predictive models from high dimensional longitudinal data, reasoning with federated knowledge bases, detecting algorithmic bias, big data analytics, analysis and prediction of protein-protein, protein-RNA, and protein-DNA interfaces and interactions, social network analytics, health informatics, secrecy-preserving query answering, representing and reasoning about preferences, and causal inference from complex, e.g., relational, data, large language models, diffusion models, and meta analysis. Honavar has been active in fostering national and international scientific collaborations in Artificial Intelligence, Data Sciences, and their applications in addressing national, international, and societal priorities in accelerating science, improving health, transforming agriculture through partnerships that bring together academia, non-profits, and industry. He is also active in making the science policy case for major national research initiatives such as AI for accelerating science and AI for combating the epidemic of diseases of despair. == Honors == National Science Foundation Director's Award for Superior Accomplishment, 2013 National Science Foundation Director's Award for Collaborative Integration, 2012 Margaret Ellen White Graduate Faculty Award, Iowa State University, 2011 Outstanding Career Achievement in Research Award, College of Liberal Arts and Sciences, Iowa State University, 2008 Regents Award for Faculty Excellence, Iowa Board of Regents, 2007 Edward Frymoyer Endowed Chair in Information Sciences and Technology, Penn State College of Information Sciences and Technology, Pennsylvania State University, 2013 Senior Faculty Research Excellence Award, Penn State College of Information Sciences and Technology, Pennsylvania State University, 2016 125 People of Impact, Department of Electrical and Computer Engineering, University of Wisconsin-Madison, 2016 Sudha Murty Distinguished (Visiting) Chair of Neurocomputing and Data Science, Indian Institute of Science, 2016-2021 ACM Distinguished Member, 2018 AAAS Fellow American Association for the Advancement of Science, 2018 EAI Fellow European Alliance for Innovation, 2019 Dorothy Foehr Huck and J. Lloyd Huck Chair in Biomedical Data Sciences and Artificial Intelligence, Pennsylvania State University, 2021
AI Code-review Tools: Free vs Paid (2026)
Comparing the best AI code-review tool? An AI code-review tool is software that uses machine learning to help you get more done — it lowers the barrier so anyone can produce professional output. Privacy matters too: check whether your data trains the model and whether a no-log or enterprise tier is available. Whether you are a beginner or a pro, the right AI code-review tool slots into your workflow and pays for itself fast. We tested the leading options and ranked them by quality, value, and ease of use.
Multimodal representation learning
Multimodal representation learning is a subfield of representation learning focused on integrating and interpreting information from different modalities, such as text, images, audio, or video, by projecting them into a shared latent space. This allows for semantically similar content across modalities to be mapped to nearby points within that space, facilitating a unified understanding of diverse data types. By automatically learning meaningful features from each modality and capturing their inter-modal relationships, multimodal representation learning enables a unified representation that enhances performance in cross-media analysis tasks such as video classification, event detection, and sentiment analysis. It also supports cross-modal retrieval and translation, including image captioning, video description, and text-to-image synthesis. == Motivation == The primary motivations for multimodal representation learning arise from the inherent nature of real-world data and the limitations of unimodal approaches. Since multimodal data offers complementary and supplementary information about an object or event from different perspectives, it is more informative than relying on a single modality. A key motivation is to narrow the heterogeneity gap that exists between different modalities by projecting their features into a shared semantic subspace. This allows semantically similar content across modalities to be represented by similar vectors, facilitating the understanding of relationships and correlations between them. Multimodal representation learning aims to leverage the unique information provided by each modality to achieve a more comprehensive and accurate understanding of concepts. These unified representations are crucial for improving performance in various cross-media analysis tasks such as video classification, event detection, and sentiment analysis. They also enable cross-modal retrieval, allowing users to search and retrieve content across different modalities. Additionally, it facilitates cross-modal translation, where information can be converted from one modality to another, as seen in applications like image captioning and text-to-image synthesis. The abundance of ubiquitous multimodal data in real-world applications, including understudied areas like healthcare, finance, and human-computer interaction (HCI), further motivates the development of effective multimodal representation learning techniques. == Approaches and methods == === Canonical-correlation analysis based methods === Canonical-correlation analysis (CCA) was first introduced in 1936 by Harold Hotelling and is a fundamental approach for multimodal learning. CCA aims to find linear relationships between two sets of variables. Given two data matrices X ∈ R n × p {\displaystyle X\in \mathbb {R} ^{n\times p}} and Y ∈ R n × q {\displaystyle Y\in \mathbb {R} ^{n\times q}} representing different modalities, CCA finds projection vectors w x ∈ R p {\displaystyle w_{x}\in \mathbb {R} ^{p}} and w y ∈ R q {\displaystyle w_{y}\in \mathbb {R} ^{q}} that maximizes the correlation between the projected variables: ρ = max w x , w y w x ⊤ Σ x y w y w x ⊤ Σ x x w x w y ⊤ Σ y y w y {\displaystyle \rho =\max _{w_{x},w_{y}}{\frac {w_{x}^{\top }\Sigma _{xy}w_{y}}{{\sqrt {w_{x}^{\top }\Sigma _{xx}w_{x}}}{\sqrt {w_{y}^{\top }\Sigma _{yy}w_{y}}}}}} such that Σ x x {\displaystyle \Sigma _{xx}} and Σ y y {\displaystyle \Sigma _{yy}} are the within-modality covariance matrices, and Σ x y {\displaystyle \Sigma _{xy}} is the between-modality covariance matrix. However, standard CCA is limited by its linearity, which led to the development of nonlinear extensions, such as kernel CCA and deep CCA. ==== Kernel CCA ==== Kernel canonical correlation analysis (KCCA) extends traditional CCA to capture nonlinear relationships between modalities by implicitly mapping the data into high dimensional feature spaces using kernel functions. Given kernel functions K x {\displaystyle K_{x}} and K y {\displaystyle K_{y}} with corresponding Gram matrices K x ∈ R n × n {\displaystyle K_{x}\in \mathbb {R} ^{n\times n}} and K y ∈ R n × n {\displaystyle K_{y}\in \mathbb {R} ^{n\times n}} , KCCA seeks coefficients α {\displaystyle \alpha } and β {\displaystyle \beta } that maximize: ρ = max α , β α ⊤ K x K y β α ⊤ K x 2 α β ⊤ K y 2 β {\displaystyle \rho =\max _{\alpha ,\beta }{\frac {\alpha ^{\top }K_{x}Ky\beta }{{\sqrt {\alpha ^{\top }K_{x}^{2}\alpha }}{\sqrt {\beta ^{\top }K_{y}^{2}\beta }}}}} To prevent overfitting, regularization terms are typically added, resulting in: ρ = max α , β α T K x K y β α T ( K x 2 + λ x K x ) α β T ( K y 2 + λ y K y ) β {\displaystyle \rho =\max _{\alpha ,\beta }{\frac {\alpha ^{T}K_{x}K_{y}\beta }{{\sqrt {\alpha ^{T}\left(K_{x}^{2}+\lambda _{x}K_{x}\right)\alpha }}{\sqrt {\;\beta ^{T}\left(K_{y}^{2}+\lambda _{y}K_{y}\right)\beta }}}}} where λ x {\displaystyle \lambda _{x}} and λ y {\displaystyle \lambda _{y}} are regularization parameters. KCCA has proven effective for tasks such as cross-modal retrieval and semantic analysis, though it faces computational challenges with large datasets due to its O ( n 2 ) {\displaystyle O(n^{2})} memory requirement for sorting kernel matrices. KCCA was proposed independently by several researchers. ==== Deep CCA ==== Deep canonical correlation analysis (DCCA), introduced in 2013, employs neural networks to learn nonlinear transformations for maximizing the correlation between modalities. DCCA uses separate neural networks f x {\displaystyle f_{x}} and f y {\displaystyle f_{y}} for each modality to transform the original data before applying CCA: max W x , W y , θ x , θ y corr ( f x ( X ; θ x ) , f y ( Y ; θ y ) ) {\displaystyle \max _{W_{x},W_{y},\theta _{x},\theta _{y}}\operatorname {corr} \left(f_{x}(X;\theta _{x}),f_{y}(Y;\theta _{y})\right)} where θ x {\displaystyle \theta _{x}} and θ y {\displaystyle \theta _{y}} represent the parameters of the neural networks, and W x {\displaystyle W_{x}} and W y {\displaystyle W_{y}} are the CCA projection matrices. The correlation objective is computed as: corr ( H x , H y ) = tr ( T − 1 / 2 H x T H y S − 1 / 2 ) {\displaystyle \operatorname {corr} (H_{x},H_{y})=\operatorname {tr} \left(T^{-1/2}H_{x}^{T}H_{y}S^{-1/2}\right)} where H x = f x ( X ) {\displaystyle H_{x}=f_{x}(X)} and H y = f y ( Y ) {\displaystyle H_{y}=f_{y}(Y)} are the network outputs, T = H x T H x + r x I {\displaystyle T=H_{x}^{T}H_{x}+r_{x}I} , S = H y T H y + r y I {\displaystyle S=H_{y}^{T}H_{y}+r_{y}I} and r x , r y {\displaystyle r_{x},r_{y}} are the regularization parameters. DCCA overcomes the limitations of linear CCA and kernel CCA by learning complex nonlinear relationships while maintaining computational efficiency for large datasets through mini-batch optimization. === Graph-based methods === Graph-based approaches for multimodal representation learning leverage graph structure to model relationships between entities across different modalities. These methods typically represent each modality as a graph and then learn embedding that preserve cross-modal similarities, enabling more effective joint representation of heterogeneous data. One such method is cross-modal graph neural networks (CMGNNs) that extend traditional graph neural networks (GNNs) to handle data from multiple modalities by constructing graphs that capture both intra-modal and inter-modal relationships. These networks model interactions across modalities by representing them as nodes and their relationships as edges. Other graph-based methods include Probabilistic Graphical Models (PGMs) such as deep belief networks (DBN) and deep Boltzmann machines (DBM). These models can learn a joint representation across modalities, for instance, a multimodal DBN achieves this by adding a shared restricted Boltzmann Machine (RBM) hidden layer on top of modality-specific DBNs. Additionally, the structure of data in some domains like Human-Computer Interaction (HCI), such as the view hierarchy of app screens, can potentially be modeled using graph-like structures. The field of graph representation learning is also relevant, with ongoing progress in developing evaluation benchmarks. === Diffusion maps === Another set of methods relevant to multimodal representation learning are based on diffusion maps and their extensions to handle multiple modalities. ==== Multi-view diffusion maps ==== Multi-view diffusion maps address the challenge of achieving multi-view dimensionality reduction by effectively utilizing the availability of multiple views to extract a coherent low-dimensional representation of the data. The core idea is to exploit both the intrinsic relations within each view and the mutual relations between the different views, defining a cross-view model where a random walk process implicitly hops between objects in different views. A multi-view kernel matrix is constructed by combining these relations, defining a cross-view diffusion process and associ
Nondeterministic finite automaton
In automata theory, a finite-state machine is called a deterministic finite automaton (DFA), if each of its transitions is uniquely determined by its source state and input symbol, and reading an input symbol is required for each state transition. A nondeterministic finite automaton (NFA), or nondeterministic finite-state machine, does not need to obey these restrictions. In particular, every DFA is also an NFA. Sometimes the term NFA is used in a narrower sense, referring to an NFA that is not a DFA, but not in this article. Using the subset construction algorithm, each NFA can be translated to an equivalent DFA; i.e., a DFA recognizing the same formal language. Like DFAs, NFAs only recognize regular languages. NFAs were introduced in 1959 by Michael O. Rabin and Dana Scott, who also showed their equivalence to DFAs. NFAs are used in the implementation of regular expressions: Thompson's construction is an algorithm for compiling a regular expression to an NFA that can efficiently perform pattern matching on strings. Conversely, Kleene's algorithm can be used to convert an NFA into a regular expression (whose size is generally exponential in the input automaton). NFAs have been generalized in multiple ways, e.g., nondeterministic finite automata with ε-moves, finite-state transducers, pushdown automata, alternating automata, ω-automata, and probabilistic automata. Besides the DFAs, other known special cases of NFAs are unambiguous finite automata (UFA) and self-verifying finite automata (SVFA). == Informal introduction == There are at least two equivalent ways to describe the behavior of an NFA. The first way makes use of the nondeterminism in the name of an NFA. For each input symbol, the NFA transitions to a new state until all input symbols have been consumed. In each step, the automaton nondeterministically "chooses" one of the applicable transitions. If there exists at least one "lucky run", i.e. some sequence of choices leading to an accepting state after completely consuming the input, it is accepted. Otherwise, i.e. if no choice sequence at all can consume all the input and lead to an accepting state, the input is rejected. In the second way, the NFA consumes a string of input symbols, one by one. In each step, whenever two or more transitions are applicable, it "clones" itself into appropriately many copies, each one following a different transition. If no transition is applicable, the current copy is in a dead end, and it "dies". If, after consuming the complete input, any of the copies is in an accept state, the input is accepted, else, it is rejected. == Formal definition == For a more elementary introduction of the formal definition, see automata theory. === Automaton === An NFA is represented formally by a 5-tuple, ( Q , Σ , δ , q 0 , F ) {\displaystyle (Q,\Sigma ,\delta ,q_{0},F)} , consisting of a finite set of states Q {\displaystyle Q} , a finite set of input symbols called the alphabet Σ {\displaystyle \Sigma } , a transition function δ {\displaystyle \delta } : Q × Σ → P ( Q ) {\displaystyle Q\times \Sigma \rightarrow {\mathcal {P}}(Q)} , an initial (or start) state q 0 ∈ Q {\displaystyle q_{0}\in Q} , and a set of accepting (or final) states F ⊆ Q {\displaystyle F\subseteq Q} . Here, P ( Q ) {\displaystyle {\mathcal {P}}(Q)} denotes the power set of Q {\displaystyle Q} . === Recognized language === Given an NFA M = ( Q , Σ , δ , q 0 , F ) {\displaystyle M=(Q,\Sigma ,\delta ,q_{0},F)} , its recognized language is denoted by L ( M ) {\displaystyle L(M)} , and is defined as the set of all strings over the alphabet Σ {\displaystyle \Sigma } that are accepted by M {\displaystyle M} . Loosely corresponding to the above informal explanations, there are several equivalent formal definitions of a string w = a 1 a 2 . . . a n {\displaystyle w=a_{1}a_{2}...a_{n}} being accepted by M {\displaystyle M} : w {\displaystyle w} is accepted if a sequence of states, r 0 , r 1 , . . . , r n {\displaystyle r_{0},r_{1},...,r_{n}} , exists in Q {\displaystyle Q} such that: r 0 = q 0 {\displaystyle r_{0}=q_{0}} r i + 1 ∈ δ ( r i , a i + 1 ) {\displaystyle r_{i+1}\in \delta (r_{i},a_{i+1})} , for i = 0 , … , n − 1 {\displaystyle i=0,\ldots ,n-1} r n ∈ F {\displaystyle r_{n}\in F} . In words, the first condition says that the machine starts in the start state q 0 {\displaystyle q_{0}} . The second condition says that given each character of string w {\displaystyle w} , the machine will transition from state to state according to the transition function δ {\displaystyle \delta } . The last condition says that the machine accepts w {\displaystyle w} if the last input of w {\displaystyle w} causes the machine to halt in one of the accepting states. In order for w {\displaystyle w} to be accepted by M {\displaystyle M} , it is not required that every state sequence ends in an accepting state, it is sufficient if one does. Otherwise, i.e. if it is impossible at all to get from q 0 {\displaystyle q_{0}} to a state from F {\displaystyle F} by following w {\displaystyle w} , it is said that the automaton rejects the string. The set of strings M {\displaystyle M} accepts is the language recognized by M {\displaystyle M} and this language is denoted by L ( M ) {\displaystyle L(M)} . Alternatively, w {\displaystyle w} is accepted if δ ∗ ( q 0 , w ) ∩ F ≠ ∅ {\displaystyle \delta ^{}(q_{0},w)\cap F\not =\emptyset } , where δ ∗ : Q × Σ ∗ → P ( Q ) {\displaystyle \delta ^{}:Q\times \Sigma ^{}\rightarrow {\mathcal {P}}(Q)} is defined recursively by: δ ∗ ( r , ε ) = { r } {\displaystyle \delta ^{}(r,\varepsilon )=\{r\}} where ε {\displaystyle \varepsilon } is the empty string, and δ ∗ ( r , x a ) = ⋃ r ′ ∈ δ ∗ ( r , x ) δ ( r ′ , a ) {\displaystyle \delta ^{}(r,xa)=\bigcup _{r'\in \delta ^{}(r,x)}\delta (r',a)} for all x ∈ Σ ∗ , a ∈ Σ {\displaystyle x\in \Sigma ^{},a\in \Sigma } . In words, δ ∗ ( r , x ) {\displaystyle \delta ^{}(r,x)} is the set of all states reachable from state r {\displaystyle r} by consuming the string x {\displaystyle x} . The string w {\displaystyle w} is accepted if some accepting state in F {\displaystyle F} can be reached from the start state q 0 {\displaystyle q_{0}} by consuming w {\displaystyle w} . === Initial state === The above automaton definition uses a single initial state, which is not necessary. Sometimes, NFAs are defined with a set of initial states. There is an easy construction that translates an NFA with multiple initial states to an NFA with a single initial state, which provides a convenient notation. == Example == The following automaton M, with a binary alphabet, determines if the input ends with a 1. Let M = ( { p , q } , { 0 , 1 } , δ , p , { q } ) {\displaystyle M=(\{p,q\},\{0,1\},\delta ,p,\{q\})} where the transition function δ {\displaystyle \delta } can be defined by this state transition table (cf. upper left picture): State Input 0 1 p { p } { p , q } q ∅ ∅ {\displaystyle {\begin{array}{|c|cc|}{\bcancel {{}_{\text{State}}\quad {}^{\text{Input}}}}&0&1\\\hline p&\{p\}&\{p,q\}\\q&\emptyset &\emptyset \end{array}}} Since the set δ ( p , 1 ) {\displaystyle \delta (p,1)} contains more than one state, M is nondeterministic. The language of M can be described by the regular language given by the regular expression (0|1)1. All possible state sequences for the input string "1011" are shown in the lower picture. The string is accepted by M since one state sequence satisfies the above definition; it does not matter that other sequences fail to do so. The picture can be interpreted in a couple of ways: In terms of the above "lucky-run" explanation, each path in the picture denotes a sequence of choices of M. In terms of the "cloning" explanation, each vertical column shows all clones of M at a given point in time, multiple arrows emanating from a node indicate cloning, a node without emanating arrows indicating the "death" of a clone. The feasibility to read the same picture in two ways also indicates the equivalence of both above explanations. Considering the first of the above formal definitions, "1011" is accepted since when reading it M may traverse the state sequence ⟨ r 0 , r 1 , r 2 , r 3 , r 4 ⟩ = ⟨ p , p , p , p , q ⟩ {\displaystyle \langle r_{0},r_{1},r_{2},r_{3},r_{4}\rangle =\langle p,p,p,p,q\rangle } , which satisfies conditions 1 to 3. Concerning the second formal definition, bottom-up computation shows that δ ∗ ( p , ε ) = { p } {\displaystyle \delta ^{}(p,\varepsilon )=\{p\}} , hence δ ∗ ( p , 1 ) = δ ( p , 1 ) = { p , q } {\displaystyle \delta ^{}(p,1)=\delta (p,1)=\{p,q\}} , hence δ ∗ ( p , 10 ) = δ ( p , 0 ) ∪ δ ( q , 0 ) = { p } ∪ { } {\displaystyle \delta ^{}(p,10)=\delta (p,0)\cup \delta (q,0)=\{p\}\cup \{\}} , hence δ ∗ ( p , 101 ) = δ ( p , 1 ) = { p , q } {\displaystyle \delta ^{}(p,101)=\delta (p,1)=\{p,q\}} , and hence δ ∗ ( p , 1011 ) = δ ( p , 1 ) ∪ δ ( q , 1 ) = { p , q } ∪ { } {\displaystyle \delta ^{}(p,1011)=\delta (p,1)\cup \delta (q,1)=\{p,q\}\cup \{\}} ; since that set is