The softmax function, also known as softargmax or normalized exponential function, converts a tuple of K real numbers into a probability distribution over K possible outcomes. It is a generalization of the logistic function to multiple dimensions, and is used in multinomial logistic regression. The softmax function is often used as the last activation function of a neural network to normalize the output of a network to a probability distribution over predicted output classes. == Definition == The softmax function takes as input a tuple z of K real numbers, and normalizes it into a probability distribution consisting of K probabilities proportional to the exponentials of the input numbers. That is, prior to applying softmax, some tuple components could be negative, or greater than one; and might not sum to 1; but after applying softmax, each component will be in the interval ( 0 , 1 ) {\displaystyle (0,1)} , and the components will add up to 1, so that they can be interpreted as probabilities. Furthermore, the larger input components will correspond to larger probabilities. Formally, the standard (unit) softmax function σ : R K → ( 0 , 1 ) K {\displaystyle \sigma :\mathbb {R} ^{K}\to (0,1)^{K}} , where K > 1 {\displaystyle K>1} , takes a tuple z = ( z 1 , … , z K ) ∈ R K {\displaystyle \mathbf {z} =(z_{1},\dotsc ,z_{K})\in \mathbb {R} ^{K}} and computes each component of vector σ ( z ) ∈ ( 0 , 1 ) K {\displaystyle \sigma (\mathbf {z} )\in (0,1)^{K}} with σ ( z ) i = e z i ∑ j = 1 K e z j . {\displaystyle \sigma (\mathbf {z} )_{i}={\frac {e^{z_{i}}}{\sum _{j=1}^{K}e^{z_{j}}}}\,.} In words, the softmax applies the standard exponential function to each element z i {\displaystyle z_{i}} of the input tuple z {\displaystyle \mathbf {z} } (consisting of K {\displaystyle K} real numbers), and normalizes these values by dividing by the sum of all these exponentials. The normalization ensures that the sum of the components of the output vector σ ( z ) {\displaystyle \sigma (\mathbf {z} )} is 1. The term "softmax" derives from the amplifying effects of the exponential on any maxima in the input tuple. For example, the standard softmax of ( 1 , 2 , 8 ) {\displaystyle (1,2,8)} is approximately ( 0.001 , 0.002 , 0.997 ) {\displaystyle (0.001,0.002,0.997)} , which amounts to assigning almost all of the total unit weight in the result to the position of the tuple's maximal element (of 8). In general, instead of e a different base b > 0 can be used. As above, if b > 1 then larger input components will result in larger output probabilities, and increasing the value of b will create probability distributions that are more concentrated around the positions of the largest input values. Conversely, if 0 < b < 1 then smaller input components will result in larger output probabilities, and decreasing the value of b will create probability distributions that are more concentrated around the positions of the smallest input values. Writing b = e β {\displaystyle b=e^{\beta }} or b = e − β {\displaystyle b=e^{-\beta }} (for real β) yields the expressions: σ ( z ) i = e β z i ∑ j = 1 K e β z j or σ ( z ) i = e − β z i ∑ j = 1 K e − β z j for i = 1 , … , K . {\displaystyle \sigma (\mathbf {z} )_{i}={\frac {e^{\beta z_{i}}}{\sum _{j=1}^{K}e^{\beta z_{j}}}}{\text{ or }}\sigma (\mathbf {z} )_{i}={\frac {e^{-\beta z_{i}}}{\sum _{j=1}^{K}e^{-\beta z_{j}}}}{\text{ for }}i=1,\dotsc ,K.} A value proportional to the reciprocal of β is sometimes referred to as the temperature: β = 1 / k T {\textstyle \beta =1/kT} , where k is typically 1 or the Boltzmann constant and T is the temperature. A higher temperature results in a more uniform output distribution (i.e. with higher entropy; it is "more random"), while a lower temperature results in a sharper output distribution, with one value dominating. In some fields, the base is fixed, corresponding to a fixed scale, while in others the parameter β (or T) is varied. The softmax function is a multiple-variable generalization of the logistic function. == Interpretations == === Smooth arg max === The Softmax function is a smooth approximation to the arg max function: the function whose value is the index of a tuple's largest element. The name "softmax" may be misleading. Softmax is not a smooth maximum (that is, a smooth approximation to the maximum function). The term "softmax" is also used for the closely related LogSumExp function, which is a smooth maximum. For this reason, some prefer the more accurate term "softargmax", though the term "softmax" is conventional in machine learning. This section uses the term "softargmax" for clarity. Formally, instead of considering the arg max as a function with categorical output 1 , … , n {\displaystyle 1,\dots ,n} (corresponding to the index), consider the arg max function with one-hot representation of the output (assuming there is a unique maximum arg): a r g m a x ( z 1 , … , z n ) = ( y 1 , … , y n ) = ( 0 , … , 0 , 1 , 0 , … , 0 ) , {\displaystyle \operatorname {arg\,max} (z_{1},\,\dots ,\,z_{n})=(y_{1},\,\dots ,\,y_{n})=(0,\,\dots ,\,0,\,1,\,0,\,\dots ,\,0),} where the output coordinate y i = 1 {\displaystyle y_{i}=1} if and only if i {\displaystyle i} is the arg max of ( z 1 , … , z n ) {\displaystyle (z_{1},\dots ,z_{n})} , meaning z i {\displaystyle z_{i}} is the unique maximum value of ( z 1 , … , z n ) {\displaystyle (z_{1},\,\dots ,\,z_{n})} . For example, in this encoding a r g m a x ( 1 , 5 , 10 ) = ( 0 , 0 , 1 ) , {\displaystyle \operatorname {arg\,max} (1,5,10)=(0,0,1),} since the third argument is the maximum. This can be generalized to multiple arg max values (multiple equal z i {\displaystyle z_{i}} being the maximum) by dividing the 1 between all max args; formally 1/k where k is the number of arguments assuming the maximum. For example, a r g m a x ( 1 , 5 , 5 ) = ( 0 , 1 / 2 , 1 / 2 ) , {\displaystyle \operatorname {arg\,max} (1,\,5,\,5)=(0,\,1/2,\,1/2),} since the second and third argument are both the maximum. In case all arguments are equal, this is simply a r g m a x ( z , … , z ) = ( 1 / n , … , 1 / n ) . {\displaystyle \operatorname {arg\,max} (z,\dots ,z)=(1/n,\dots ,1/n).} Points z with multiple arg max values are singular points (or singularities, and form the singular set) – these are the points where arg max is discontinuous (with a jump discontinuity) – while points with a single arg max are known as non-singular or regular points. With the last expression given in the introduction, softargmax is now a smooth approximation of arg max: as β → ∞ {\displaystyle \beta \to \infty } , softargmax converges to arg max. There are various notions of convergence of a function; softargmax converges to arg max pointwise, meaning for each fixed input z as β → ∞ {\displaystyle \beta \to \infty } , σ β ( z ) → a r g m a x ( z ) . {\displaystyle \sigma _{\beta }(\mathbf {z} )\to \operatorname {arg\,max} (\mathbf {z} ).} However, softargmax does not converge uniformly to arg max, meaning intuitively that different points converge at different rates, and may converge arbitrarily slowly. In fact, softargmax is continuous, but arg max is not continuous at the singular set where two coordinates are equal, while the uniform limit of continuous functions is continuous. The reason it fails to converge uniformly is that for inputs where two coordinates are almost equal (and one is the maximum), the arg max is the index of one or the other, so a small change in input yields a large change in output. For example, σ β ( 1 , 1.0001 ) → ( 0 , 1 ) , {\displaystyle \sigma _{\beta }(1,\,1.0001)\to (0,1),} but σ β ( 1 , 0.9999 ) → ( 1 , 0 ) , {\displaystyle \sigma _{\beta }(1,\,0.9999)\to (1,\,0),} and σ β ( 1 , 1 ) = 1 / 2 {\displaystyle \sigma _{\beta }(1,\,1)=1/2} for all inputs: the closer the points are to the singular set ( x , x ) {\displaystyle (x,x)} , the slower they converge. However, softargmax does converge compactly on the non-singular set. Conversely, as β → − ∞ {\displaystyle \beta \to -\infty } , softargmax converges to arg min in the same way, where here the singular set is points with two arg min values. In the language of tropical analysis, the softmax is a deformation or "quantization" of arg max and arg min, corresponding to using the log semiring instead of the max-plus semiring (respectively min-plus semiring), and recovering the arg max or arg min by taking the limit is called "tropicalization" or "dequantization". It is also the case that, for any fixed β, if one input z i {\displaystyle z_{i}} is much larger than the others relative to the temperature, T = 1 / β {\displaystyle T=1/\beta } , the output is approximately the arg max. For example, a difference of 10 is large relative to a temperature of 1: σ ( 0 , 10 ) := σ 1 ( 0 , 10 ) = ( 1 / ( 1 + e 10 ) , e 10 / ( 1 + e 10 ) ) ≈ ( 0.00005 , 0.99995 ) {\displaystyle \sigma (0,\,10):=\sigma _{1}(0,\,10)=\left(1/\left(1+e^{10}\right),\,e^{10}/\left(1+e^{10}\right)\right)\approx (0.00005
Flo (app)
Flo is a period-tracking app that provides menstrual cycle, ovulation and pregnancy tracking as well as perimenopause symptom tracking that was developed by Flo Health, Inc. It has over 380 million downloads worldwide and over 70 million monthly active users as of November 2024. In mid-2024, it reached unicorn status, and became Europe’s first femtech unicorn. The company has been accused of sharing users' sensitive health data with third parties without consent and misleading its users about data practices. == History == Flo Health, Inc. was co-founded in 2015 by Dmitry and Yuri Gurski, in Belarus. Their backgrounds helped build the first version of the software having experience in other fitness and health apps. Dmitry serves as the company's CEO. The company's development hubs are in London, Amsterdam and Vilnius. In 2016, the company raised $1 million in seed round funding from Flint Capital and Haxus Venture Fund. In 2017, Flo received an investment of $5 million from Flint Capital and model Natalia Vodianova with Vodianova helping develop an awareness campaign for the company. In 2018, Flo received an investment of $6 million from Mangrove Capital Partners, with participation from Flint Capital and Haxus, giving the company a valuation of $200 million. In mid-2019, Flo received an additional investment of $7.5 million led by Founders Fund. In 2020, the Federal Trade Commission alleged that Flo had misled users about its handling of health information to third parties including Google, Facebook, AppsFlyer, and Flurry since 2016. These allegations followed a 2019 report by The Wall Street Journal in reference to Facebook. The company reached a settlement in 2021 and was required to notify users of how their personal information was shared and obtain permission before any further information was shared. The agreement also required that Flo to undertake an independent privacy audit which it completed in March 2022. In early September 2021, Flo announced it closed $50M in a Series B financing, bringing the total capital raised to $65 million and company valuation to $800M led by VNV Global and Target Global. In March 2024, the Supreme Court of British Columbia certified a class action suit against Flo for sharing intimate data with Facebook and other third parties without user knowledge. In July 2024, Flo announced it raised more than $200M in Series C financing from General Atlantic bringing its valuation beyond $1 billion. As of November 2024, the app had over 380 million downloads world wide, and over 70 million monthly active users. In 2025, Flo adopted a data intelligence platform from Databricks to power its analytics and AI features, allowing users personalized cycle predictions. In 2025, a class action lawsuit in California was settled for $56 million with Flo paying $8 million and Google paying $48 million. == Features and privacy == Flo was initially created as a period and ovulation tracking application. It now provides reminders of upcoming menstrual cycles and a place to record various other health symptoms such as contraceptive methods, vaginal discharge (leukorrhea), water intake, pains, mood swings, and sexual activity. The application is available on iOS and Android. Flo is free to download and the free basic version gives you access to period and ovulation tracking and predictions, symptom tracking, cycle history, and anonymous mode. In Pregnancy mode, the app provides tracking features and educational material for pregnancy. In October 2023, Flo launched Flo for Partners, a feature that allows users to share their Flo data with their partner. In September 2022, as a response to Roe v. Wade being overturned, Flo sped up the release of a feature called "Anonymous Mode". Flo said this mode allows users to access the app without any personal identifiers such as name, email address, or technical identifiers being associated with their health data. Flo said it uses a technology called Oblivious HTTP to help protect user privacy in Anonymous Mode. == Recognition == Flo was named to Bloomberg’s Top 25 UK Startups to Watch for 2024. Flo's Anonymous Mode feature was recognized on both Fast Company's World Changing Ideas 2023 and TIME's Best Inventions List 2023. Flo is a CES 2019 Innovation Awards Honoree in the Software and Mobile Applications category.
Sumio Watanabe
Sumio Watanabe (渡辺 澄夫, Watanabe Sumio; born 1959) is a Japanese mathematician and engineer working in probability theory, applied algebraic geometry and Bayesian statistics. He is currently a professor at Tokyo Institute of Technology in the Department of Computational Intelligence and Systems Science. He is the author of the text, Algebraic Geometry and Statistical Learning Theory, which proposes a generalization of Fisher's regular statistical theory to singular statistical models. == Books == Mathematical Theory of Bayesian Statistics, CRC Press, 2018, ISBN 9781482238068 Algebraic Geometry and Statistical Learning Theory, Cambridge University Press, 2009.
AI Text-to-video Tools: Free vs Paid (2026)
Curious about the best AI text-to-video tool? An AI text-to-video tool is software that uses machine learning to help you get more done — it combines speed, accuracy, and an interface that just works. Hands-on testing shows real-world results vary, so a short free trial is the smartest way to decide. Whether you are a beginner or a pro, the right AI text-to-video tool slots into your workflow and pays for itself fast. Read on for hands-on impressions, pricing tiers, and the standout features that matter.
Judea Pearl
Judea Pearl (Hebrew: יהודה פרל; born September 4, 1936) is an Israeli-American electrical engineer, computer scientist and philosopher, best known for championing the probabilistic approach to artificial intelligence and the development of Bayesian networks (see the article on belief propagation). He is also credited for developing a theory of causal and counterfactual inference based on structural models (see article on causality). In 2011, the Association for Computing Machinery (ACM) awarded Pearl with the Turing Award, the highest distinction in computer science, "for fundamental contributions to artificial intelligence through the development of a calculus for probabilistic and causal reasoning". He is the author of several books, including the technical Causality: Models, Reasoning and Inference, and The Book of Why, a book on causality aimed at the general public. Judea Pearl is the father of journalist Daniel Pearl, who was kidnapped and murdered by terrorists in Pakistan connected with Al-Qaeda and the International Islamic Front in 2002. == Biography == Judea Pearl was born in Tel Aviv, British Mandate for Palestine, in 1936 to Eliezer and Tova Pearl, who were Polish Jewish immigrants, grew up in Bnei Brak. His grandfather Chaim Pearl was one of Bnei Brak's founders. He is a descendant of Menachem Mendel of Kotzk on his mother's side. After serving in the Israel Defense Forces and joining a kibbutz, Pearl decided to study engineering in 1956. He received a B.S. in electrical engineering from the Technion 1960. That same year, he emigrated to the United States and pursued graduate studies. He received an M.S. in electrical engineering from the Newark College of Engineering (now New Jersey Institute of Technology) in 1961, and went on to receive an M.S. in physics from Rutgers University and a PhD in electrical engineering from the Polytechnic Institute of Brooklyn (now the New York University Tandon School of Engineering) in 1965. He worked at RCA Research Laboratories (now SRI International) in Princeton, New Jersey on superconductive parametric amplifiers and storage devices and at Electronic Memories, Inc., on advanced memory systems. When semiconductors "wiped out" Pearl's work, as he later expressed it, he joined UCLA's School of Engineering in 1970 and started work on probabilistic artificial intelligence. He is one of the founding editors of the Journal of Causal Inference. Pearl is currently a professor of computer science and statistics and director of the Cognitive Systems Laboratory at UCLA. He and his wife, Ruth, had three children. In addition, as of 2011, he is a member of the International Advisory Board of NGO Monitor. Former Israeli Chief Rabbi, Rabbi Yisrael Meir Lau, partnered with Judea Pearl in the documentary With My Whole Broken Heart. == Murder of Daniel Pearl == In 2002, his son, Daniel Pearl, a journalist working for the Wall Street Journal was kidnapped and murdered in Pakistan, leading Judea and the other members of the family and friends to create the Daniel Pearl Foundation. On the seventh anniversary of Daniel's death, Judea wrote an article in the Wall Street Journal titled Daniel Pearl and the Normalization of Evil: When will our luminaries stop making excuses for terror?. Emeritus Chief Rabbi Jonathan Sacks quoted Judea Pearl's beliefs in a lesson on Judaism: "I asked Judea Pearl, father of the murdered journalist Daniel Pearl, why he was working for reconciliation between Jews and Muslims...he replied with heartbreaking lucidity, 'Hate killed my son. Therefore I am determined to fight hate.'" == Views == On his religious views, Pearl states that he is a "practicing disbeliever." He is very connected to Jewish traditions such as holidays and kiddush on Friday night. Pearl sits on the NGO Monitor international advisory board, a right-wing organization based in Jerusalem that reports on non-governmental organization activity from a pro-Israel perspective. == Research == Pearl is credited for "laying the foundations of modern artificial intelligence, so computer systems can process uncertainty and relate causes to effects." He is one of the pioneers of Bayesian networks and the probabilistic approach to artificial intelligence, and one of the first to mathematize causal modeling in the empirical sciences. His work is also intended as a high-level cognitive model. He is interested in the philosophy of science, knowledge representation, nonstandard logics, and learning. Pearl is described as "one of the giants in the field of artificial intelligence" by UCLA computer science professor Richard E. Korf. His work on causality has "revolutionized the understanding of causality in statistics, psychology, medicine and the social sciences" according to the Association for Computing Machinery. === Notable contributions === A summary of Pearl's scientific contributions is available in a chronological account authored by Stuart J. Russell (2012). An annotated bibliography of Pearl's contributions was compiled by the ACM in 2012. A video describing Pearl's major contributions to AI is available here. Pearl's opinion pieces, touching on Jewish identity, the war on terrorism, and the Middle East conflict can be accessed here. === Books === Heuristics, Addison-Wesley, 1984 Probabilistic Reasoning in Intelligent Systems, Morgan-Kaufmann, 1988 Pearl, Judea (2000). Causality: Models, Reasoning, and Inference. Cambridge University Press. I Am Jewish: Personal Reflections Inspired by the Last Words of Daniel Pearl, Jewish Lights, 2004. (Winner of a 2004 National Jewish Book Award) Causal Inference in Statistics: A Primer, (with Madelyn Glymour and Nicholas Jewell), Wiley, 2016. ISBN 978-1-119-18684-7 A previous survey: Causal inference in statistics: An overview, Statistics Surveys, 3:96–146, 2009. Pearl, Judea; Dana Mackenzie (2018). "The Book of Why: The New Science of Cause and Effect". Science. 361 (6405): 855. Bibcode:2018Sci...361..855.. doi:10.1126/science.aau9731. === Awards ===
Clef (app)
Clef was a San Francisco-based technology company, known for developing a mobile app that created a two-factor authentication for websites. It allowed users to access sites with a single login password management service which stores encrypted passwords in private accounts. It had a standard verification method that requires access to data on the mobile phone to confirm the user's identity. The application required a Wi-Fi or mobile network, and the user could log in by scanning the computer screen with their phone. == History == Clef was founded in 2013 by Mark Hudnall, B. Byrne and Jesse Pollak. It raised $1.6 million in seed funding in November 2014. Clef integrated with many websites and applications, including WordPress. On March 17, 2017, Clef announced they would no longer support the plugin after June 6, 2017; Clef was acquired by Authy, another 2FA service, which later got acquired by Twilio.
Neuroph
Neuroph is an object-oriented artificial neural network framework written in Java. It can be used to create and train neural networks in Java programs. Neuroph provides Java class library as well as GUI tool easyNeurons for creating and training neural networks. It is an open-source project hosted at SourceForge under the Apache License. Versions before 2.4 were licensed under LGPL 3, from this version the license is Apache 2.0 License. == Features == Neuroph's core classes correspond to basic neural network concepts like artificial neuron, neuron layer, neuron connections, weight, transfer function, input function, learning rule etc. Neuroph supports common neural network architectures such as Multilayer perceptron with Backpropagation, Kohonen and Hopfield networks. All these classes can be extended and customized to create custom neural networks and learning rules. Neuroph has built-in support for image recognition.