Region-based Convolutional Neural Networks (R-CNN) are a family of machine learning models for computer vision, and specifically object detection and localization. The original goal of R-CNN was to take an input image and produce a set of bounding boxes as output, where each bounding box contains an object and also the category (e.g. car or pedestrian) of the object. In general, R-CNN architectures perform selective search over feature maps outputted by a CNN. R-CNN has been extended to perform other computer vision tasks, such as: tracking objects from a drone-mounted camera, locating text in an image, and enabling object detection in Google Lens. Mask R-CNN is also one of seven tasks in the MLPerf Training Benchmark, which is a competition to speed up the training of neural networks. == History == The following covers some of the versions of R-CNN that have been developed. November 2013: R-CNN. April 2015: Fast R-CNN. June 2015: Faster R-CNN. March 2017: Mask R-CNN. December 2017: Cascade R-CNN is trained with increasing Intersection over Union (IoU, also known as the Jaccard index) thresholds, making each stage more selective against nearby false positives. June 2019: Mesh R-CNN adds the ability to generate a 3D mesh from a 2D image. == Architecture == For review articles see. === Selective search === Given an image (or an image-like feature map), selective search (also called Hierarchical Grouping) first segments the image by the algorithm in (Felzenszwalb and Huttenlocher, 2004), then performs the following: Input: (colour) image Output: Set of object location hypotheses L Segment image into initial regions R = {r1, ..., rn} using Felzenszwalb and Huttenlocher (2004) Initialise similarity set S = ∅ foreach Neighbouring region pair (ri, rj) do Calculate similarity s(ri, rj) S = S ∪ s(ri, rj) while S ≠ ∅ do Get highest similarity s(ri, rj) = max(S) Merge corresponding regions rt = ri ∪ rj Remove similarities regarding ri: S = S \ s(ri, r∗) Remove similarities regarding rj: S = S \ s(r∗, rj) Calculate similarity set St between rt and its neighbours S = S ∪ St R = R ∪ rt Extract object location boxes L from all regions in R === R-CNN === With R-CNN, prediction follows a two-step process. A preprocessing selective search step generates a large set of candidate objects (typically as many as 2000), known as regions of interest (ROI). These are forwarded to a CNN, which predicts an object class score and bounding box estimate, independently for each ROI. Importantly, the ROIs are heavily filtered to remove excess candidates. This is achieved using two mechanism. Filtering begins by removing ROIs assigned to the background category. This is a specialized category, which is scored by the CNN alongside other categories. An unfortunate reality is that remaining ROIs typically suffer from heavy duplication. Namely, multiple ROIs that cover same objects in the image are all assigned non-background categories. This is resolved by a heuristic non-maximum suppression (NMS) step. === Fast R-CNN === While the original R-CNN independently computed the neural network features on each of as many as two thousand regions of interest, Fast R-CNN runs the neural network once on the whole image. At the end of the network is a ROIPooling module, which slices out each ROI from the network's output tensor, reshapes it, and classifies it. As in the original R-CNN, the Fast R-CNN uses selective search to generate its region proposals. === Faster R-CNN === While Fast R-CNN used selective search to generate ROIs, Faster R-CNN integrates the ROI generation into the neural network itself. === Mask R-CNN === While previous versions of R-CNN focused on object detections, Mask R-CNN adds instance segmentation. Mask R-CNN also replaced ROIPooling with a new method called ROIAlign, which can represent fractions of a pixel.
Active learning (machine learning)
Active learning is a special case of machine learning in which a learning algorithm can interactively query a human user (or some other information source) to label new data points with the desired outputs. The human user must possess expertise in the problem domain, including the ability to consult authoritative sources when necessary. In statistics literature, it is sometimes also called optimal experimental design. The information source is also called teacher or oracle. There are situations in which unlabeled data is abundant but manual labeling is expensive. In such a scenario, learning algorithms can actively query the teacher for labels. Since the learner chooses the examples, the number of examples to learn a concept can often be much lower than the number required in normal supervised learning. However, there is a risk that the algorithm is overwhelmed by uninformative examples. Recent developments are dedicated to multi-label active learning, hybrid active learning and active learning in a single-pass (on-line) context, combining concepts from the field of machine learning (e.g. conflict and ignorance) with adaptive, incremental learning policies in the field of online machine learning. Using active learning allows for faster development of a machine learning algorithm, when comparative updates would require a quantum or super computer. Large-scale active learning projects may benefit from crowdsourcing frameworks such as Amazon Mechanical Turk that include many humans in the active learning loop. == Definitions == Let T be the total set of all data under consideration. For example, in a protein engineering problem, T would include all proteins that are known to have a certain interesting activity and all additional proteins that one might want to test for that activity. During each iteration, i, T is broken up into three subsets T K , i {\displaystyle \mathbf {T} _{K,i}} : Data points where the label is known. T U , i {\displaystyle \mathbf {T} _{U,i}} : Data points where the label is unknown. T C , i {\displaystyle \mathbf {T} _{C,i}} : A subset of TU,i that is chosen to be labeled. Most of the current research in active learning involves the best method to choose the data points for TC,i. == Scenarios == Pool-based sampling: In this approach, which is the most well known scenario, the learning algorithm attempts to evaluate the entire dataset before selecting data points (instances) for labeling. It is often initially trained on a fully labeled subset of the data using a machine-learning method such as logistic regression or SVM that yields class-membership probabilities for individual data instances. The candidate instances are those for which the prediction is most ambiguous. Instances are drawn from the entire data pool and assigned a confidence score, a measurement of how well the learner "understands" the data. The system then selects the instances for which it is the least confident and queries the teacher for the labels. The theoretical drawback of pool-based sampling is that it is memory-intensive and is therefore limited in its capacity to handle enormous datasets, but in practice, the rate-limiting factor is that the teacher is typically a (fatiguable) human expert who must be paid for their effort, rather than computer memory. Stream-based selective sampling: Here, each consecutive unlabeled instance is examined one at a time with the machine evaluating the informativeness of each item against its query parameters. The learner decides for itself whether to assign a label or query the teacher for each datapoint. As contrasted with Pool-based sampling, the obvious drawback of stream-based methods is that the learning algorithm does not have sufficient information, early in the process, to make a sound assign-label-vs ask-teacher decision, and it does not capitalize as efficiently on the presence of already labeled data. Therefore, the teacher is likely to spend more effort in supplying labels than with the pool-based approach. Membership query synthesis: This is where the learner generates synthetic data from an underlying natural distribution. For example, if the dataset are pictures of humans and animals, the learner could send a clipped image of a leg to the teacher and query if this appendage belongs to an animal or human. This is particularly useful if the dataset is small. The challenge here, as with all synthetic-data-generation efforts, is in ensuring that the synthetic data is consistent in terms of meeting the constraints on real data. As the number of variables/features in the input data increase, and strong dependencies between variables exist, it becomes increasingly difficult to generate synthetic data with sufficient fidelity. For example, to create a synthetic data set for human laboratory-test values, the sum of the various white blood cell (WBC) components in a white blood cell differential must equal 100, since the component numbers are really percentages. Similarly, the enzymes alanine transaminase (ALT) and aspartate transaminase (AST) measure liver function (though AST is also produced by other tissues, e.g., lung, pancreas) A synthetic data point with AST at the lower limit of normal range (8–33 units/L) with an ALT several times above normal range (4–35 units/L) in a simulated chronically ill patient would be physiologically impossible. == Query strategies == Algorithms for determining which data points should be labeled can be organized into a number of different categories, based upon their purpose: Balance exploration and exploitation: the choice of examples to label is seen as a dilemma between the exploration and the exploitation over the data space representation. This strategy manages this compromise by modelling the active learning problem as a contextual bandit problem. For example, Bouneffouf et al. propose a sequential algorithm named Active Thompson Sampling (ATS), which, in each round, assigns a sampling distribution on the pool, samples one point from this distribution, and queries the oracle for this sample point label. Expected model change: label those points that would most change the current model. Expected error reduction: label those points that would most reduce the model's generalization error. Exponentiated Gradient Exploration for Active Learning: In this paper, the author proposes a sequential algorithm named exponentiated gradient (EG)-active that can improve any active learning algorithm by an optimal random exploration. Uncertainty sampling: label those points for which the current model is least certain as to what the correct output should be. Query by committee: a variety of models are trained on the current labeled data, and vote on the output for unlabeled data; label those points for which the "committee" disagrees the most Querying from diverse subspaces or partitions: When the underlying model is a forest of trees, the leaf nodes might represent (overlapping) partitions of the original feature space. This offers the possibility of selecting instances from non-overlapping or minimally overlapping partitions for labeling. Variance reduction: label those points that would minimize output variance, which is one of the components of error. Conformal prediction: predicts that a new data point will have a label similar to old data points in some specified way and degree of the similarity within the old examples is used to estimate the confidence in the prediction. Mismatch-first farthest-traversal: The primary selection criterion is the prediction mismatch between the current model and nearest-neighbour prediction. It targets on wrongly predicted data points. The second selection criterion is the distance to previously selected data, the farthest first. It aims at optimizing the diversity of selected data. User-centered labeling strategies: Learning is accomplished by applying dimensionality reduction to graphs and figures like scatter plots. Then the user is asked to label the compiled data (categorical, numerical, relevance scores, relation between two instances). A wide variety of algorithms have been studied that fall into these categories. While the traditional AL strategies can achieve remarkable performance, it is often challenging to predict in advance which strategy is the most suitable in a particular situation. In recent years, meta-learning algorithms have been gaining in popularity. Some of them have been proposed to tackle the problem of learning AL strategies instead of relying on manually designed strategies. A benchmark which compares 'meta-learning approaches to active learning' to 'traditional heuristic-based Active Learning' may give intuitions if 'Learning active learning' is at the crossroads == Minimum marginal hyperplane == Some active learning algorithms are built upon support-vector machines (SVMs) and exploit the structure of the SVM to determine which data points to label. Such methods usually calculate the margin, W, of each u
Best AI Clip Makers in 2026
Trying to pick the best AI clip maker? An AI clip maker 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 clip maker slots into your workflow and pays for itself fast. Read on for hands-on impressions, pricing tiers, and the standout features that matter.
AI Art Generators Reviews: What Actually Works in 2026
Comparing the best AI art generator? An AI art generator 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 art generator slots into your workflow and pays for itself fast. Below we compare features, pricing, and real output so you can choose with confidence.
Evaluation of machine translation
Various methods for the evaluation for machine translation have been employed. This article focuses on the evaluation of the output of machine translation, rather than on performance or usability evaluation. == Round-trip translation == A typical way for lay people to assess machine translation quality is to translate from a source language to a target language and back to the source language with the same engine. Though intuitively this may seem like a good method of evaluation, it has been shown that round-trip translation is a "poor predictor of quality". The reason why it is such a poor predictor of quality is reasonably intuitive. A round-trip translation is not testing one system, but two systems: the language pair of the engine for translating into the target language, and the language pair translating back from the target language. Consider the following examples of round-trip translation performed from English to Italian and Portuguese from Somers (2005): In the first example, where the text is translated into Italian then back into English—the English text is significantly garbled, but the Italian is a serviceable translation. In the second example, the text translated back into English is perfect, but the Portuguese translation is meaningless; the program thought "tit" was a reference to a tit (bird), which was intended for a "tat", a word it did not understand. While round-trip translation may be useful to generate a "surplus of fun," the methodology is deficient for serious study of machine translation quality. == Human evaluation == This section covers two of the large scale evaluation studies that have had significant impact on the field—the ALPAC 1966 study and the ARPA study. === Automatic Language Processing Advisory Committee (ALPAC) === One of the constituent parts of the ALPAC report was a study comparing different levels of human translation with machine translation output, using human subjects as judges. The human judges were specially trained for the purpose. The evaluation study compared an MT system translating from Russian into English with human translators, on two variables. The variables studied were "intelligibility" and "fidelity". Intelligibility was a measure of how "understandable" the sentence was, and was measured on a scale of 1–9. Fidelity was a measure of how much information the translated sentence retained compared to the original, and was measured on a scale of 0–9. Each point on the scale was associated with a textual description. For example, 3 on the intelligibility scale was described as "Generally unintelligible; it tends to read like nonsense but, with a considerable amount of reflection and study, one can at least hypothesize the idea intended by the sentence". Intelligibility was measured without reference to the original, while fidelity was measured indirectly. The translated sentence was presented, and after reading it and absorbing the content, the original sentence was presented. The judges were asked to rate the original sentence on informativeness. So, the more informative the original sentence, the lower the quality of the translation. The study showed that the variables were highly correlated when the human judgment was averaged per sentence. The variation among raters was small, but the researchers recommended that at the very least, three or four raters should be used. The evaluation methodology managed to separate translations by humans from translations by machines with ease. The study concluded that, "highly reliable assessments can be made of the quality of human and machine translations". === Advanced Research Projects Agency (ARPA) === As part of the Human Language Technologies Program, the Advanced Research Projects Agency (ARPA) created a methodology to evaluate machine translation systems, and continues to perform evaluations based on this methodology. The evaluation programme was instigated in 1991, and continues to this day. Details of the programme can be found in White et al. (1994) and White (1995). The evaluation programme involved testing several systems based on different theoretical approaches; statistical, rule-based and human-assisted. A number of methods for the evaluation of the output from these systems were tested in 1992 and the most recent suitable methods were selected for inclusion in the programmes for subsequent years. The methods were; comprehension evaluation, quality panel evaluation, and evaluation based on adequacy and fluency. Comprehension evaluation aimed to directly compare systems based on the results from multiple choice comprehension tests, as in Church et al. (1993). The texts chosen were a set of articles in English on the subject of financial news. These articles were translated by professional translators into a series of language pairs, and then translated back into English using the machine translation systems. It was decided that this was not adequate for a standalone method of comparing systems and as such abandoned due to issues with the modification of meaning in the process of translating from English. The idea of quality panel evaluation was to submit translations to a panel of expert native English speakers who were professional translators and get them to evaluate them. The evaluations were done on the basis of a metric, modelled on a standard US government metric used to rate human translations. This was good from the point of view that the metric was "externally motivated", since it was not specifically developed for machine translation. However, the quality panel evaluation was very difficult to set up logistically, as it necessitated having a number of experts together in one place for a week or more, and furthermore for them to reach consensus. This method was also abandoned. Along with a modified form of the comprehension evaluation (re-styled as informativeness evaluation), the most popular method was to obtain ratings from monolingual judges for segments of a document. The judges were presented with a segment, and asked to rate it for two variables, adequacy and fluency. Adequacy is a rating of how much information is transferred between the original and the translation, and fluency is a rating of how good the English is. This technique was found to cover the relevant parts of the quality panel evaluation, while at the same time being easier to deploy, as it didn't require expert judgment. Measuring systems based on adequacy and fluency, along with informativeness is now the standard methodology for the ARPA evaluation program. == Automatic evaluation == In the context of this article, a metric is a measurement. A metric that evaluates machine translation output represents the quality of the output. The quality of a translation is inherently subjective, there is no objective or quantifiable "good." Therefore, any metric must assign quality scores so they correlate with the human judgment of quality. That is, a metric should score highly translations that humans score highly, and give low scores to those humans give low scores. Human judgment is the benchmark for assessing automatic metrics, as humans are the end-users of any translation output. The measure of evaluation for metrics is correlation with human judgment. This is generally done at two levels, at the sentence level, where scores are calculated by the metric for a set of translated sentences, and then correlated against human judgment for the same sentences. And at the corpus level, where scores over the sentences are aggregated for both human judgments and metric judgments, and these aggregate scores are then correlated. Figures for correlation at the sentence level are rarely reported, although Banerjee et al. (2005) do give correlation figures that show that, at least for their metric, sentence-level correlation is substantially worse than corpus level correlation. While not widely reported, it has been noted that the genre, or domain, of a text has an effect on the correlation obtained when using metrics. Coughlin (2003) reports that comparing the candidate text against a single reference translation does not adversely affect the correlation of metrics when working in a restricted domain text. Even if a metric correlates well with human judgment in one study on one corpus, this successful correlation may not carry over to another corpus. Good metric performance, across text types or domains, is important for the reusability of the metric. A metric that only works for text in a specific domain is useful, but less useful than one that works across many domains—because creating a new metric for every new evaluation or domain is undesirable. Another important factor in the usefulness of an evaluation metric is to have a good correlation, even when working with small amounts of data, that is candidate sentences and reference translations. Turian et al. (2003) point out that, "Any MT evaluation measure is less reliable on shorter translations", and
PowerBuilder
PowerBuilder is an integrated development environment owned by SAP since the acquisition of Sybase in 2010. On July 5, 2016, SAP and Appeon entered into an agreement whereby Appeon, an independent company, would be responsible for developing, selling, and supporting PowerBuilder. Over the years, PowerBuilder has been updated with new standards. In 2010, a major upgrade of PowerBuilder was released to provide support for the Microsoft .NET Framework. In 2014, support was added for OData, dockable windows, and 64-bit native applications. In 2019 support was added for rapidly creating RESTful Web APIs and non-visual .NET assemblies using the C# language and the .NET Core framework. And PowerScript client app development was revamped with new UI technologies and cloud architecture. In 2025 the IDE was revamped with new code editor and ultra-fast compiler. Appeon has been releasing new features every 6-12 month cycles, which per the product roadmap focus on four key focus areas: sustaining core features, modernizing application UI, improving developer productivity, and incorporating more Cloud technology. == Features == PowerBuilder has a native data-handling component called a DataWindow, which can be used to create, edit, and display data from a database. This object gives the programmer a number of tools for specifying and controlling user interface appearance and behavior, and also provides simplified access to database content and JSON or XML from Web services. To some extent, the DataWindow frees the programmer from considering the differences between Database Management Systems from different vendors. DataWindow can display data using multiple presentation styles and can connect to various data sources. == Usage == PowerBuilder is used primarily for building business-oriented CRUD applications. Although new software products are rarely built with PowerBuilder, many client-server ERP products and line-of-business applications built in the late 1980s to early 2000s with PowerBuilder still provide core database functions for large enterprises in government, higher education, manufacturing, insurance, banking, energy, and telecommunications. == History == === Early history === PowerBuilder originated from Computer Solutions Inc. (CSI), a software consulting firm founded in 1974 by Mitchell Kertzman in Massachusetts. CSI developed GrowthPower, an MRP II software package with integrated financial modules released in 1981, which ran exclusively on the HP 3000 platform and achieved over 1,000 customer installations at its peak. In the late 1980s, as demand increased for graphical user interfaces amid the rise of Microsoft Windows, Kertzman partnered with Dave Litwack, former executive vice president of product development at Cullinet Software (acquired by Computer Associates in 1989). Litwack joined the company in 1988 as head of research and development to develop a client/server GUI tool, leading to its rebranding as Powersoft Corporation in 1990. PowerBuilder 1.0 was released in July 1991 as a rapid application development tool featuring the DataWindow and PowerScript language. Powersoft went public on February 3, 1993, with shares closing at $38 from an initial $20 price. Sybase announced its acquisition of Powersoft on November 15, 1994, in a stock swap valued at approximately $940 million; the merger closed on February 14, 1995, at a revised value of about $904 million due to Sybase's stock fluctuations. === Recent history === In December 2013 SAP announced the new version going directly to number 15 and released a beta version. Key features included support for the .NET Framework v4.5, SQL Server 2012, Oracle 12, Windows 8, OData and Dockable Windows. SAP later released this as version 12.6. On May 31, 2019, PowerBuilder 2019 was released by Appeon. This release supports C# development. It provides a new C# IDE, .NET data access objects, C# migration solution, Web API client, and UI themes. On April 3, 2020, PowerBuilder 2019 R2 was launched by Appeon. This release includes a first-ever PowerScript-to-C# code converter, which can automatically migrate 80-95% of PowerBuilder business logic and DataWindows to C#. Interoperability between PowerScript and .NET programming languages is also now supported. Many existing features have also been enhanced. On January 22, 2021, PowerBuilder 2019 R3 was launched by Appeon. This release provides a groundbreaking new app deployment technology called PowerClient, which securely automates the installation and update of client apps over HTTPS. C# Web API development has been greatly enhanced with asynchronous programming and support for Amazon Aurora and Azure cloud databases. Aside from many other new features, PowerBuilder 2019 R3 is a long-term support (LTS) version that replaces previous LTS versions On August 6, 2021, PowerBuilder 2021 was launched by Appeon. The Cloud deployment capability of the PowerBuilder 2021 IDE, in conjunction with the matching PowerServer 2021 runtime, was revamped, bringing PowerBuilder up-to-date with the latest .NET technologies. The presentation layer now executes PowerScript natively on Windows devices. The middle-tier has been rebuilt around REST API standard with a pure .NET Core implementation. A new CI/CD utility that integrates with Git/SVN and Jenkins, witch compiles all PowerBuilder projects using the command-line interface, was added alongside other features. On September 4, 2022, PowerBuilder 2022 was launched by Appeon. This release brings enhancements to the productivity of developing both client/server & installable cloud apps and more security measures to safeguard your apps. It includes many new features, including Windows 11 support, introducing time-saving functionalities to the IDE, such as Tabbed Code Editor, Jump to Objects, and Quick Code Search, and supports the latest HTTP/2 and TLS 1.3 protocols and two-way TLS authentication. On August 4, 2023, PowerBuilder 2022 R2 was launched by Appeon. This release introduces a range of new features aimed at helping developers build powerful, feature-rich, and secure client/server and installable cloud apps more efficiently, including tabbed windows, fillable PDFs, and SMTP client. On January 8, 2024, PowerBuilder 2022 R3 was launched by Appeon. This release is a long-term support version. Features previously released in earlier releases have been enhanced and/or corrected. On May 7, 2025, PowerBuilder 2025 was launched by Appeon. This release delivers a revamped IDE that boosts developer productivity throughout the SLDC—from writing and extending code to debugging, automating builds, and deploying applications. It features a new-generation code editor, ultra-fast compiler, automatic REST API creation, faster GIT operations, and codeless UI modernization features. == Features == PowerBuilder is an object-oriented programming language. Nearly all of the visual and non-visual objects support inheritance, polymorphism, and encapsulation. The programmer may utilize a common code framework such as PowerBuilder Foundation Classes, also known as PFC, to inherit objects from and leverage pre-existing code. The DataWindow is the key component (and selling point) of PowerBuilder. The DataWindow offers a visual SQL painter which supports outer joins, unions and subquery operations. It can convert SQL to visual representation and back, so the developer can use native SQL if desired. DataWindow updates are automatic — it produces the proper SQL at runtime based on the DBMS to which the user is currently connected. This feature makes it easier for developers who are not experienced with SQL. The DataWindow also has the built-in ability to both retrieve data and update data via stored procedures or REST Web APIs as well as import/export JSON data. The RESTClient object introduced in PowerBuilder 2017 facilitates bridging the DataWindow with REST Web APIs and requiring minimal coding. === RDBMS interfaces === PowerBuilder offers native interfaces to all major databases, as well as ODBC and OLE-DB, in the Enterprise version. There are many connectivity options that allow performance monitoring and tuning, such as: Integrated security Tracing of all SQL Isolation level Password expiration dialog Blocking factor Number of SQL statements to cache Use connection pool Thread safety Trace ODBC API calls Due to the information about the database schema (such as primary key information) that are stored in PowerBuilder's data dictionary, the code required to implement data display and browsing is greatly simplified, because the dictionary information allows generation of the appropriate SQL behind the scenes. PowerBuilder supports the following ways of interacting with a database: DataWindow this is the simplest approach, relying on automatically generated SQL. Embedded SQL Embedded SQL supports SELECT, INSERT, UPDATE, DELETE and cursors. This option is used when the developer desires more control than is available with the
Heng Ji
Heng Ji is a computer scientist who works on information extraction and natural language processing. She is well known for her work on joined named entity recognition and relation extraction, as well as for her work on cross-document event extraction. She has been coordinating the popular NIST TAC Knowledge Base Population task since 2010. She has been recognised as one of AI's 10 to watch by IEEE Intelligent Systems in 2013, and has won multiple awards, including a NSF Career Award in 2009, Google Research awards in 2009 and 2014, and an IBM Watson Faculty Award in 2012. == Education == Heng Ji obtained a Bachelor's and master's degree in Computational Linguistics from Tsinghua University. She subsequently obtained a MSc, then PhD in Computer Science from New York University in 2008 under the supervision of Ralph Grishman. Her PhD thesis was on the topic of information extraction, with a particular focus on joint training of multiple components in the information extraction pipeline, as well as cross-lingual learning. == Career == Upon graduating with a PhD from New York University, Ji took up a position as assistant professor at Queens College, City University of New York, where she founded the BLENDER Lab, which focuses on research on cross-lingual, cross-documents, cross-media information extraction and fusion. In 2013, she joined Rensselaer Polytechnic Institute as an Edward P. Hamilton Development Chair and Tenured associate professor in Computer Science. Since 2019, she has been a full professor at the University of Illinois at Urbana–Champaign, as well as an Amazon Scholar. == Research == Heng Ji works in the area of natural language processing, machine learning and information extraction. She has published over 300 peer-reviewed research papers. Her work is published in the proceedings of computer science conferences, including the Annual Meeting of the Association for Computational Linguistics, The Web Conference, and the ACM Conference on Knowledge Discovery and Data Mining (KDD). Ji is a leading researcher in information extraction, having coordinated the popular NIST TAC Knowledge Base Population shared task since 2010. She is most recognised for her work on modelling interactions between subtasks in information extraction, which was also the topic of her PhD thesis, and for her work on event detection using cross-document signals. == Selected honors and distinctions == 2009 NSF Career Award 2009 Google Research Award 2012 IBM Watson Faculty Award 2013 IEEE AI's 10 to Watch 2014 Google Research Award 2016 World Economic Forum, 'Young Scientist' 2017 World Economic Forum, 'Young Scientist' 2020 Annual Meeting of the Association for Computational Linguistics, best demonstration paper