Patent Grant
12248883
The subject matter described herein relates to techniques for identifying or otherwise characterizing a prompt injection attack on an artificial intelligence (AI) model, such as a large language model, using a prompt injection classifier.
Machine learning (ML) algorithms and models, such as large language models, ingest large amounts of data and use pattern recognition and other techniques to make predictions and adjustments based on that data. These models have attack surfaces that can be vulnerable to cyberattacks in which adversaries attempt to manipulate or modify model behavior. These attacks can act to corrupt input data so as to make outputs unreliable or incorrect. By modifying or otherwise manipulating the input of a model, an attacker can modify an output of an application or process for malicious purposes including bypassing security measures resulting in data leakage, unauthorized system access, and the solicitation of unlawful or otherwise restricted information.
In a first aspect, an analysis engine receives data characterizing a prompt for ingestion by a generative artificial intelligence (GenAI) model. The analysis engine, using a prompt injection classifier, determines whether the prompt comprises malicious content, elicits malicious actions or other undesired GenAI model behavior. Data characterizing the determination is later provided to a consuming application or process.
The data characterizing the prompt can be vectorized to result in one or more vectors. One or more embeddings can be generated based on the one or more vectors (which can have a lower dimensionality than the corresponding vectors). The prompt injection classifier can use these embeddings when making the determination. In some variations, the prompt injection classifier uses both features corresponding to the embeddings as well as other features derived from other information.
The GenAI model can take various forms including a large language model.
The consuming application or process can take various actions based on the determination. For example, the consuming application or process can allow the prompt to be input into the GenAI model upon a determination that the prompt does not comprise malicious content and/or it can prevent the prompt from being input into the GenAI model upon a determination that the prompt comprises or elicits malicious content. The consuming application or process can flag (e.g., generate an alert, label, etc.) the prompt as being malicious for quality assurance upon a determination that the prompt comprises or elicits malicious content. The consuming application or process can modify the prompt to be benign upon a determination that the prompt comprises or elicits malicious content and can cause the modified prompt to be ingested by the GenAI model. The consuming application or process can block an entity identified by one or more of an internet protocol (IP) address, a media access control (MAC) address, session, or other identifier of a requester of the prompt upon a determination that the prompt comprises or elicits malicious content. The consuming application or process can cause subsequent prompts from the IP address of the requester of the prompt to be modified upon a determination that the prompt comprises or elicits malicious content and can cause the modified prompt to be ingested by the GenAI model. In cases in which an IP address is hosting multiple clients, a combination of the MAC address and the IP address can be blocked or, in other implementations, the HTTP session can be terminated/restarted as well as blocking the MAC address and the IP address.
In an interrelated aspect, the prompt injection classifier can be a multi-class model. With such an arrangement, an analysis engine receives data characterizing a prompt for ingestion by a generative artificial intelligence (GenAI) model. The analysis engine uses the prompt injection classifier to determine a category for the prompt which is indicative of whether the prompt comprises or elicits malicious content (or otherwise causes the GenAI model to behave in an undesired manner. Data characterizing such a determination can be provided to a consuming application or process.
The categories can take varying forms. As one example, the category can specify a threat severity for the prompt (malicious, suspicious, unknown, or benign, etc.). In other variations, the category can specify a type of prompt injection attack. The prompt injection attack types can include, for example, one or more of: a direct task deflection attack, a special case attack, a context continuation attack, a context termination attack, a syntactic transformation attack, an encryption attack, a text redirection attack, as well as other types of prompt injection attacks. In these variations, one or more remediation actions can be initiated which are tailored to the specified type of prompt injection attack (as determined by the prompt injection classifier). In some cases, the prompt injection classifier can comprise a plurality of different machine learning models (e.g., an ensemble of machine learning models) in which at least some of such models are trained to categorize a different type of prompt injection attack.
The consuming application or process can take various actions based on the determined category of the prompt. For example, the consuming application or process can allow the prompt to be input into the GenAI model upon a determination that the prompt is of a category that does not comprise or elicit malicious content. The consuming application or process can prevent the prompt from being input into the GenAI model upon a determination that the prompt is of a category that comprises or elicits malicious content. The consuming application or process can flag the prompt as being malicious for quality assurance upon a determination that the prompt is of a category that comprises or elicits malicious content. The consuming application or process can modify the prompt to be benign upon a determination that the prompt is of a category that comprises or elicits malicious content and can cause the modified prompt to be ingested by the GenAI model. The consuming application or process can block an IP address of a requester of the prompt upon a determination that the prompt is of a category that comprises or elicits malicious content. The consuming application or process can cause subsequent prompts from an entity identified by an IP address, MAC address, session or other identifier of a requester of the prompt to be modified upon a determination that the prompt is of a category that comprises or elicits malicious content and can cause the modified prompt to be ingested by the GenAI model.
In a further interrelated aspect, an analysis engine receives data characterizing a prompt for ingestion by a generative artificial intelligence (GenAI) model. The analysis engine, using a prompt injection classifier, determines whether the prompt is indicative of comprising malicious content or eliciting malicious actions. The prompt is allowed to be ingested by the GenAI model if it is determined that the prompt is not indicative of comprising malicious content or eliciting malicious actions. At least one remediation action is initiated if it is determined that the prompt is indicative of comprising malicious content or eliciting malicious actions.
Non-transitory computer program products (i.e., physically embodied computer program products) are also described that comprise instructions, which when executed by one or more data processors of one or more computing systems, cause at least one data processor to perform operations herein. Similarly, computer systems are also described that may include one or more data processors and memory coupled to the one or more data processors. The memory may temporarily or permanently store instructions that cause at least one processor to perform one or more of the operations described herein. In addition, methods can be implemented by one or more data processors either within a single computing system or distributed among two or more computing systems. Such computing systems can be connected and can exchange data and/or commands or other instructions or the like via one or more connections, including but not limited to a connection over a network (e.g., the Internet, a wireless wide area network, a local area network, a wide area network, a wired network, or the like), via a direct connection between one or more of the multiple computing systems, etc.
The subject matter described herein provides many technical advantages. For example, the current subject matter can be used to identify and stop adversarial attacks on artificial intelligence models including large language models. Further, the current subject matter can provide enhanced visibility into the health and security of an enterprise's machine learning assets.
The details of one or more variations of the subject matter described herein are set forth in the accompanying drawings and the description below. Other features and advantages of the subject matter described herein will be apparent from the description and drawings, and from the claims.
Like reference symbols in the various drawings indicate like elements.
The current subject matter is directed to advanced techniques for identifying and preventing cyberattacks on advanced artificial intelligence (AI) models including large language models. In particular, the current subject matter is directed to analyzing prompts of an GenAI model to determine, using machine learning, whether they are malicious or benign, and in some variations, a particular type of prompt injection attack can be identified. Malicious as used herein can refer to actions which cause the GenAI model to respond in an undesired manner. With these classifications, remediation actions can be taken in connection with the prompt including blocking the prompt, modifying the prompt, disconnecting the requesting device, disconnecting the account, and the like.
The proxy 150 can communicate, over one or more networks, with a monitoring environment 160. The monitoring environment 160 can include one or more servers and data stores to execute an analysis engine 170. The analysis engine 170 can execute one or more of the algorithms/models described below with regard to the protection of the MLA 130.
The proxy 150 can, in some variations, relay received queries to the monitoring environment 160 prior to ingestion by the MLA 130. The proxy 150 can also or alternatively relay information which characterizes the received queries (e.g., excerpts, extracted features, metadata, etc.) to the monitoring environment 160 prior to ingestion by the MLA 130.
The analysis engine 170 can analyze the relayed queries and/or information in order to make an assessment or other determination as to whether the queries are indicative of being malicious. In some cases, a remediation engine 180 which can form part of the monitoring environment 160 (or be external such as illustrated in
The proxy 150 can, in some variations, relay outputs of the MLA to the monitoring environment 160 prior to transmission to the respective client device 110. The proxy 150 can also or alternatively relay information which characterizes the outputs (e.g., excerpts, extracted features, metadata, etc.) to the monitoring environment 160 prior to transmission to the respective client device 110.
The analysis engine 170 can analyze the relayed outputs and/or information from the MLA 130 in order to make an assessment or other determination as to whether the queries are indicative of being malicious (based on the output alone or based on combination of the input and the output). In some cases, the remediation engine 180 can, similar to the actions when the query analysis above, take one or more remediation actions in response to a determination of a query as being malicious. These remediation actions can take various forms including transmitting data to the proxy 150 which causes the output of the MLA 130 to be blocked prior to transmission to the requesting client device 110. In some cases, the remediation engine 180 can cause data to be transmitted to the proxy 150 which causes the output for transmission to the requesting client device 110 to be modified in order to be non-malicious, to remove sensitive information, and the like.
As indicated above, one or more of the analysis engines 152, 170 can include, execute, or otherwise instantiate a prompt injection classifier 192, 194 which, in some variations, is a binary classifier which can identify a prompt as being malicious or benign. In some variations, the prompt injection classifier 192, 194 can be a multi-class classifier which can characterize different aspects of a prompt such as, but not limited to, a level of trustworthiness of the prompt (e.g., malicious, suspicious, benign, etc.). In some variations, the prompt injection classifier 192, 194 can be a multi-class classifier which identifies which of a plurality of different attack types are implicated by an input prompt. Two or more of these prompt injection classifiers 192, 194 can form an ensemble of classifiers (i.e., machine learning models). The ensemble of prompt injection classifiers can be arranged such that two or more of the classifiers are executing in parallel. In other variations, the ensemble of prompt injection classifiers can be arranged such that two or more classifiers are working in sequence. For example, a binary classifier can first analyze a prompt to determine whether the prompt is malicious or benign. If the prompt is classified as being malicious, a multi-class classifier can analyze the prompt to determine a particular type of injection attack. This classification by type can be used to take remediation actions which are specifically tailored to the type of attack. Such an arrangement can also be advantageous when the multi-class classifier is more computationally expensive than the binary classifier (which avoids every prompt being analyzed by the multi-class classifier). Other arrangements can be provided with a lightweight classified being executed by the analysis engine 152 in the model environment 140 and a more computationally expensive model can be executed by the analysis engine 170 in the monitoring environment 160.
The prompt injection classifier 192, 194 can be a machine learning model such as an XGBoost classification model, a logistic regression model, an XLNet model and the like. In the case of a binary classifier, the prompt injection classifier 192, 194 can be trained using a corpus of data which can include a plurality of benign prompts that do not contain prompt injection information and a plurality of malicious prompts that contain various character strings (which can include portions of alphanumeric symbols, non-printable characters, symbols, controls, etc.) and the like which encapsulate various sorts of prompt injection. Malicious prompts in this context refer to prompts that cause the prompt injection classifier 192, 194 to exhibit undesired behavior. Benign prompts in this context can refer to prompts that do not cause the prompt injection classifier 192, 194 to exhibit undesired behavior. In some variations, the prompts forming part of the corpus can be labeled with their classification. The model training can be performed by converting the prompts into sentence embeddings which can, amongst other features, be used to train the prompt injection classifier 192, 194.
In the case of a multi-class classifier, the training corpus for the prompt injection classifier 192, 194 can include different sets of prompts for each category (i.e., severity level, type of attack, etc.) which are labeled with their category (e.g., security level, type of attack, etc.). The prompts can be transformed into sentence embeddings which can be used, amongst other features, to train the prompt injection classifier 192, 194.
The prompt injection classifier 192, 194 can be periodically retrained as new prompt injection techniques are identified and/or new remediation tools are created. Such an arrangement is advantageous in that the prompt injection classifier 192, 194 can evolve to address the continually changing threat landscape.
After the prompt injection classifier 192, 194 has been trained, the analysis engine 152, 170 can preprocess incoming prompts so that they are suitable for ingestion by the prompt injection classifier 192, 194. For example, the raw/original prompt is transformed into sentence embeddings and then input into the prompt injection classifier 192, 194 which then results in a model prediction. The model prediction for a binary classifier can predict the confidence of the prompt injection classifier. The output of the model can take varying forms including, for example, a score closer to 1 indicating that the prompt is malicious and a score closer to 0 is indicating that the prompt is benign. The model prediction for the multi-class classifiers can identify a category for the prompt (i.e., a class for which the prompt injection classifier 192, 194 has been trained).
The multi-class classifier variation of the prompt injection classifier 192, 194 can be used to identify a type of attack and, in some cases, take remedial actions which are specifically tailored to that type of attack (e.g., an attempt to obtain sensitive information or otherwise manipulate an output of the MLA 130). Example attacks include for which the prompt injection classifier 192, 194 can be trained include, but are not limited to: a direct task deflection attack, a special case attack, a context continuation attack, a context termination attack, a syntactic transformation attack, an encryption attack, a text redirection attack and the like. A direct task deflection attack can include, for example, assigning the MLA 130 a persona unrelated to its original purpose and directing it to do something is not intentionally intended to do. A special case attack can include attempts to obfuscate malicious prompts by injecting special case characters randomly or methodically, to confuse the MLA 130 to output a malicious response. A context continuation attack can include providing the MLA 130 with a single prompt or multiple prompts which follow some permutation of a pattern like: benign prompt, malicious prompt, benign prompt, continuation of malicious prompt and which, in combination, can trigger a malicious output. A context termination attack can include provoking a malicious response from the MLA 130 by providing a context and requesting the MLA 130 to essentially “fill in the blanks”. A syntactic transformation attack can include manipulation of the syntax or structure of an input to trigger or otherwise stimulate a malicious response. An encryption attack can include encrypting the prompt and tasking the MLA 130 to decrypt the prompt specifying the encryption method. A text redirection attack can include manipulating or redirecting the flow of text-based communications between users or systems. One or more of the model environment remediation engine 154, the monitoring environment remediation engine 180, or the external remediation resources 190 can take or otherwise initiate remediation activities that are specific to the type of attack and/or based on the severity classification for the prompt (e.g., malicious, highly suspicious, unknown, unable to classify, etc.). One remediation activity can be to block the IP address of the requester (i.e., the computing device initiating or otherwise relaying the prompt/input for ingestions by the MLA 130). In some cases, multiple remediation activities can be utilized such as blocking an IP address in combination with a MAC address or terminating/restarting an HTTP session while also blocking the IP and MAC addresses.
The IP address can also be used to filter (i.e., modify or otherwise redact) prompts before they are input into the MLA 130. The remediation activities can also include generating alerts (e.g., sysadmin alerts) indicating suspicious/malicious prompts. Further, the remediation activities can include capturing system/process behavior associated with suspicious/malicious prompts for analytics or other tracking purposes.
Data which characterizes the determination can then be provided, at 160, to a consuming application or process. For example, the analysis engine 152 can provide the determination to the remediation engine 154, the analysis engine 170 can provide the determination to the remediation engine 180, the analysis engine 152 can provide the determination to the remediation engine 180, the analysis engine 170 can provide the determination to the external remediation resources 190, the analysis engine 152 can provide the determination to the external remediation resources 190, and/or the determination can be transmitted to or otherwise consumed by a local or remote application or process. The analysis engine 152, 170 in this context can act as a gatekeeper to the GenAI model by sending information to a consuming application or process which results in preventing prompts deemed to be malicious from being input and allowing prompts deemed to be safe to be input. In some cases, the consuming application or process flags the prompt as being malicious for quality assurance upon a determination that the prompt comprises malicious content. In some cases, it may be desirable to modify a prompt (which can be performed by the consuming application or process) so that it ultimately is non-malicious. For example, only portions of the prompt may be deemed malicious and such aspects can be deleted or modified prior to ingestion by the GenAI model. Other actions can be taken based on the IP address of the requester (such as blocking the prompt, blocking subsequent prompts, modifying subsequent prompts, etc.). Such an arrangement still provides the attacker with an output/response thereby potentially masking the fact that the system identified the response as being malicious.
Various implementations of the subject matter described herein may be realized in digital electronic circuitry, integrated circuitry, specially designed ASICs (application specific integrated circuits), computer hardware, firmware, software, and/or combinations thereof. These various implementations may include implementation in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor (e.g., CPU, GPU, etc.), which may be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.
These computer programs (also known as programs, software, software applications or code) include machine instructions for a programmable processor, and may be implemented in a high-level procedural and/or object-oriented programming language, and/or in assembly/machine language. As used herein, the term “machine-readable medium” refers to any computer program product, apparatus and/or device (e.g., magnetic discs, optical disks, memory, Programmable Logic Devices (PLDs)) used to provide machine instructions and/or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term “machine-readable signal” refers to any signal used to provide machine instructions and/or data to a programmable processor.
To provide for interaction with a user, the subject matter described herein may be implemented on a computing device having a display device (e.g., a LED or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and an input device (e.g., mouse, trackball, touchpad, touchscreen, etc.) by which the user may provide input to the computing device. Other kinds of devices may be used to provide for interaction with a user as well; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic, speech, or tactile input.
The subject matter described herein may be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a client computer having a graphical user interface or a Web browser through which a user may interact with an implementation of the subject matter described herein), or any combination of such back-end, middleware, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (“LAN”), a wide area network (“WAN”), and the Internet.
The computing system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.
In the descriptions above and in the claims, phrases such as “at least one of” or “one or more of” may occur followed by a conjunctive list of elements or features. The term “and/or” may also occur in a list of two or more elements or features. Unless otherwise implicitly or explicitly contradicted by the context in which it is used, such a phrase is intended to mean any of the listed elements or features individually or any of the recited elements or features in combination with any of the other recited elements or features. For example, the phrases “at least one of A and B;” “one or more of A and B;” and “A and/or B” are each intended to mean “A alone, B alone, or A and B together.” A similar interpretation is also intended for lists including three or more items. For example, the phrases “at least one of A, B, and C;” “one or more of A, B, and C;” and “A, B, and/or C” are each intended to mean “A alone, B alone, C alone, A and B together, A and C together, B and C together, or A and B and C together.” In addition, use of the term “based on,” above and in the claims is intended to mean, “based at least in part on,” such that an unrecited feature or element is also permissible.
The subject matter described herein can be embodied in systems, apparatus, methods, and/or articles depending on the desired configuration. The implementations set forth in the foregoing description do not represent all implementations consistent with the subject matter described herein. Instead, they are merely some examples consistent with aspects related to the described subject matter. Although a few variations have been described in detail above, other modifications or additions are possible. In particular, further features and/or variations can be provided in addition to those set forth herein. For example, the implementations described above can be directed to various combinations and subcombinations of the disclosed features and/or combinations and subcombinations of several further features disclosed above. In addition, the logic flows depicted in the accompanying figures and/or described herein do not necessarily require the particular order shown, or sequential order, to achieve desirable results. Other implementations may be within the scope of the following claims.
| Number | Name | Date | Kind |
|---|---|---|---|
| 7802298 | Hong et al. | Sep 2010 | B1 |
| 9356941 | Kislyuk et al. | May 2016 | B1 |
| 9516053 | Muddu et al. | Dec 2016 | B1 |
| 10193902 | Caspi et al. | Jan 2019 | B1 |
| 10210036 | Iyer et al. | Feb 2019 | B2 |
| 10462168 | Shibahara et al. | Oct 2019 | B2 |
| 10637884 | Apple et al. | Apr 2020 | B2 |
| 10673880 | Pratt et al. | Jun 2020 | B1 |
| 10764313 | Mushtaq | Sep 2020 | B1 |
| 10803188 | Rajput et al. | Oct 2020 | B1 |
| 11310270 | Weber et al. | Apr 2022 | B1 |
| 11483327 | Hen et al. | Oct 2022 | B2 |
| 11501101 | Ganesan et al. | Nov 2022 | B1 |
| 11551137 | Echauz et al. | Jan 2023 | B1 |
| 11601468 | Angel et al. | Mar 2023 | B2 |
| 11710067 | Harris et al. | Jul 2023 | B2 |
| 11762998 | Kuta et al. | Sep 2023 | B2 |
| 11777957 | Chen et al. | Oct 2023 | B2 |
| 11875130 | Bosnjakovic et al. | Jan 2024 | B1 |
| 11893111 | Sai et al. | Feb 2024 | B2 |
| 11893358 | Lakshmikanthan et al. | Feb 2024 | B1 |
| 11930039 | Geethakumar et al. | Mar 2024 | B1 |
| 11960514 | Taylert et al. | Apr 2024 | B1 |
| 11962546 | Hattangady et al. | Apr 2024 | B1 |
| 11971914 | Watson et al. | Apr 2024 | B1 |
| 11972333 | Horesh et al. | Apr 2024 | B1 |
| 11997059 | Su et al. | May 2024 | B1 |
| 12124592 | O'Hern et al. | Oct 2024 | B1 |
| 20100082811 | Van Der Merwe | Apr 2010 | A1 |
| 20140033307 | Schmidtler | Jan 2014 | A1 |
| 20140157415 | Abercrombie | Jun 2014 | A1 |
| 20150074392 | Boivie et al. | Mar 2015 | A1 |
| 20160344770 | Verma et al. | Nov 2016 | A1 |
| 20170154021 | Vidhani et al. | Jun 2017 | A1 |
| 20170251006 | LaRosa et al. | Aug 2017 | A1 |
| 20170331841 | Ilu et al. | Nov 2017 | A1 |
| 20180018590 | Szeto et al. | Jan 2018 | A1 |
| 20180063190 | Wright et al. | Mar 2018 | A1 |
| 20180205734 | Wing | Jul 2018 | A1 |
| 20180219888 | Apostolopoulos | Aug 2018 | A1 |
| 20180324193 | Ronen | Nov 2018 | A1 |
| 20190050564 | Pogorelik et al. | Feb 2019 | A1 |
| 20190238572 | Manadhata et al. | Aug 2019 | A1 |
| 20190260784 | Stockdale et al. | Aug 2019 | A1 |
| 20190311118 | Grafi | Oct 2019 | A1 |
| 20190392176 | Taron et al. | Dec 2019 | A1 |
| 20200019721 | Shanmugam et al. | Jan 2020 | A1 |
| 20200076771 | Maier et al. | Mar 2020 | A1 |
| 20200175094 | Palmer et al. | Jun 2020 | A1 |
| 20200219009 | Dao et al. | Jul 2020 | A1 |
| 20200233979 | Maraghoosh et al. | Jul 2020 | A1 |
| 20200285737 | Kraus et al. | Sep 2020 | A1 |
| 20200409323 | Spalt et al. | Dec 2020 | A1 |
| 20210110062 | Oliner et al. | Apr 2021 | A1 |
| 20210141897 | Seifert et al. | May 2021 | A1 |
| 20210209464 | Bala et al. | Jul 2021 | A1 |
| 20210224425 | Nasr-Azadani et al. | Jul 2021 | A1 |
| 20210319098 | Pogorelik et al. | Oct 2021 | A1 |
| 20210319784 | Le Roux et al. | Oct 2021 | A1 |
| 20210357508 | Elovici et al. | Nov 2021 | A1 |
| 20210374247 | Sultana et al. | Dec 2021 | A1 |
| 20210407051 | Pardeshi et al. | Dec 2021 | A1 |
| 20220030009 | Hasan | Jan 2022 | A1 |
| 20220058444 | Olabiyi et al. | Feb 2022 | A1 |
| 20220070195 | Sern | Mar 2022 | A1 |
| 20220083658 | Shah et al. | Mar 2022 | A1 |
| 20220114399 | Castiglione et al. | Apr 2022 | A1 |
| 20220147597 | Bhide et al. | May 2022 | A1 |
| 20220164444 | Prudkovskij | May 2022 | A1 |
| 20220166795 | Simioni et al. | May 2022 | A1 |
| 20220182410 | Tupsamudre et al. | Jun 2022 | A1 |
| 20220253464 | Sloane et al. | Aug 2022 | A1 |
| 20220269796 | Chase et al. | Aug 2022 | A1 |
| 20220309179 | Payne et al. | Sep 2022 | A1 |
| 20230008037 | Venugopal et al. | Jan 2023 | A1 |
| 20230027149 | Kuan | Jan 2023 | A1 |
| 20230049479 | Mozo Velasco et al. | Feb 2023 | A1 |
| 20230109426 | Hashimoto et al. | Apr 2023 | A1 |
| 20230148116 | Stokes et al. | May 2023 | A1 |
| 20230169397 | Smith et al. | Jun 2023 | A1 |
| 20230185912 | Sinn et al. | Jun 2023 | A1 |
| 20230185915 | Rao et al. | Jun 2023 | A1 |
| 20230208869 | Bisht et al. | Jun 2023 | A1 |
| 20230229960 | Zhu et al. | Jul 2023 | A1 |
| 20230252178 | Ruelke et al. | Aug 2023 | A1 |
| 20230259787 | David et al. | Aug 2023 | A1 |
| 20230269263 | Yarabolu | Aug 2023 | A1 |
| 20230274003 | Liu et al. | Aug 2023 | A1 |
| 20230289604 | Chan et al. | Sep 2023 | A1 |
| 20230351143 | Kutt | Nov 2023 | A1 |
| 20230359903 | Cefalu | Nov 2023 | A1 |
| 20230385085 | Singh | Nov 2023 | A1 |
| 20230388324 | Thompson | Nov 2023 | A1 |
| 20240022585 | Burns et al. | Jan 2024 | A1 |
| 20240039948 | Koc et al. | Feb 2024 | A1 |
| 20240045959 | Marson et al. | Feb 2024 | A1 |
| 20240078337 | Kamyshenko et al. | Mar 2024 | A1 |
| 20240080333 | Burns et al. | Mar 2024 | A1 |
| 20240126611 | Phanishayee et al. | Apr 2024 | A1 |
| 20240160902 | Padgett et al. | May 2024 | A1 |
| 20240289628 | Parmar et al. | Aug 2024 | A1 |
| 20240296315 | Singh et al. | Sep 2024 | A1 |
| 20240386103 | Clement et al. | Nov 2024 | A1 |
| Entry |
|---|
| Arora, Rakhi, Rishi Gupta, and Pradeep Yadav. “Utilizing Ensemble Learning to enhance the detection of Malicious URLs in the Twitter dataset.” 2024 4th International Conference on Innovative Practices in Technology and Management (ICIPTM). IEEE, 2024. (Year: 2024). |
| Gupta, Maanak, et al. “From chatgpt to threatgpt: Impact of generative ai in cybersecurity and privacy.” IEEE Access (2023). (Year: 2023). |
| Morozov et al., 2019, “Unsupervised Neural Quantization for Compressed-Domain Similarity Search,” International Conference on Computer Vision (ICCV) 2019 (11 pages). |
| Rijthoven et al., 2021, “HookNet: Multi-resolution convulational neural networks for semantic segmentation in histopathology whole-slide images,” Medical Imange Analysis 68:1-10. |
