Abstract: A data processing system implements receiving a plurality of natural language prompts for a language model; storing the prompts in a privacy protecting datastore in which the prompts are inaccessible from outside of the privacy protecting datastore; analyzing the prompts stored to generate embedding vectors representing the plurality of natural language prompts; storing the embedding vectors in the privacy protecting datastore; analyzing the embedding vectors using a clustering algorithm to generate clusters of related embedding vectors in the privacy protecting datastore; generating, for each cluster, theme information summarizing the embedding vectors of the cluster; and storing the theme information associated with each cluster of the plurality of clusters in a second datastore outside of the privacy protecting datastore, the second datastore being accessible for querying the data stored therein.

Abstract: A data processing system implements obtaining a plurality of unlabeled user prompts including an unknown mixture of malicious prompts and benign prompts; analyzing each unlabeled user prompt using a multimodal vision language model to obtain embeddings representing each unlabeled user prompt; analyzing the embeddings to determine representation of each unlabeled user prompt of the plurality of unlabeled user prompts in a latent space; determining a first region of the latent space associated with benign user prompts and a second region of the latent space associated with malicious user prompts; generating labeled training data by labeling each unlabeled user prompt of the plurality of unlabeled user prompts with an indication whether each unlabeled user prompt is a benign user prompt falling with the first region or a malicious user prompt falling within the second region; and training a prompt classifier using the labeled training data.

Abstract: A data processing system implements obtaining user prompts s that include instructions to an AI application to perform one or more tasks; storing the user prompts in a prompts datastore in a secure computing environment; analyzing the user prompts using an LLM operating within the secure computing environment to generate normalized prompts based on the user prompts; extracting first n-grams from the normalized prompts using differentially private n-gram extraction that preserves user-level privacy; generating masked normalized prompts by comparing the normalized prompts with the first n-grams and replacing, with a placeholder n-gram, n-grams of the normalized prompts that do not match an n-gram of the first n-grams; extracting second n-grams from the masked normalized prompts using the differentially private n-gram extraction that preserves user-level privacy; outputting the second n-grams from the secure computing environment; and storing the second n-grams in an anonymized prompts datastore outside of the s

Abstract: A system for development of an Artificial Intelligence (AI) model while protecting sensitive user information includes: a confidential computing environment in which original prompts to the AI model written by users are collected; a trained synthetic prompt generator to generate synthetic prompts based on the original prompts, wherein the synthetic prompt generator generates anonymized synthetic prompts without sensitive user information identifiable from the original prompts; and a developer computing environment in which the synthetic prompts are submitted to the AI model under development to generate a dataset that includes the synthetic prompts and corresponding AI model output for analysis to determine updates for the AI model while protecting the sensitive user information of actual users.

Abstract: A system for development of an Artificial Intelligence (AI) model while protecting sensitive user information includes: a confidential computing environment in which original prompts to the AI model written by users are collected; a trained synthetic prompt generator to generate synthetic prompts based on the original prompts, wherein the synthetic prompt generator generates anonymized synthetic prompts without sensitive user information identifiable from the original prompts; and a developer computing environment in which the synthetic prompts are submitted to the AI model under development to generate a dataset that includes the synthetic prompts and corresponding AI model output for analysis to determine updates for the AI model while protecting the sensitive user information of actual users.

Abstract: A data processing system implements receiving a plurality of natural language prompts for a language model; storing the prompts in a privacy protecting datastore in which the prompts are inaccessible from outside of the privacy protecting datastore; analyzing the prompts stored to generate embedding vectors representing the plurality of natural language prompts; storing the embedding vectors in the privacy protecting datastore; analyzing the embedding vectors using a clustering algorithm to generate clusters of related embedding vectors in the privacy protecting datastore; generating, for each cluster, theme information summarizing the embedding vectors of the cluster; and storing the theme information associated with each cluster of the plurality of clusters in a second datastore outside of the privacy protecting datastore, the second datastore being accessible for querying the data stored therein.

Abstract: An adjustable desk mountable workstation (90) comprising: a platform (12) adapted to support a computer keyboard thereon; and a pair of symmetrically opposed legs (91) for standing or mounting on a desk and on which the platform (12) is mounted, the legs (91) being adjustable so as to alter the height of the platform (12) above the desk, each leg (91) having outer and inner ends, wherein adjustment of the legs (91) involves concomitant pivoting without displacement at their respective outer ends and pivoting plus horizontal displacement at their respective inner ends, thereby altering the height of the platform (91) with respect to the desk.

Abstract: An adjustable desk mountable workstation (90) comprising: a platform (12) adapted to support a computer keyboard thereon; and a pair of symmetrically opposed legs (91) for standing or mounting on a desk and on which the platform (12) is mounted, the legs (91) being adjustable so as to alter the height of the platform (12) above the desk, each leg (91) having outer and inner ends, wherein adjustment of the legs (91) involves concomitant pivoting without displacement at their respective outer ends and pivoting plus horizontal displacement at their respective inner ends, thereby altering the height of the platform (91) with respect to the desk.

Abstract: In one embodiment, a human interactive proof portal 140 may use a biometric input to determine whether a user is a standard user or a malicious actor. The human interactive proof portal 140 may receive an access request 302 for an online data service 122 from a user device 110. The human interactive proof portal 140 may send a proof challenge 304 to the user device 110 for presentation to a user. The human interactive proof portal 140 may receive from the user device 110 a proof response 306 having a biometric metadata description 430 based on a biometric input from the user.

Abstract: In one embodiment, a human interactive proof portal 140 may use a biometric input to determine whether a user is a standard user or a malicious actor. The human interactive proof portal 140 may receive an access request 302 for an online data service 122 from a user device 110. The human interactive proof portal 140 may send a proof challenge 304 to the user device 110 for presentation to a user. The human interactive proof portal 140 may receive from the user device 110 a proof response 306 having a biometric metadata description 430 based on a biometric input from the user.

Abstract: In one embodiment, a human interactive proof portal 140 may use a biometric input to determine whether a user is a standard user or a malicious actor. The human interactive proof portal 140 may receive an access request 302 for an online data service 122 from a user device 110. The human interactive proof portal 140 may send a proof challenge 304 to the user device 110 for presentation to a user. The human interactive proof portal 140 may receive from the user device 110 a proof response 306 having a biometric metadata description 430 based on a biometric input from the user.

Abstract: In one embodiment, an intelligent detection system 102 may determine if a network target 108 is an adversarial site based on comparing responses to different network sources. The intelligent detection system 102 may select a test apparent network source 110 and a control apparent network source 112 from a network source pool 106. The intelligent detection system 102 may receive the test response responding to a test request from the test apparent network source 110 to a network target 108. The intelligent detection system 102 may receive the control response responding to a control request from the control apparent network source 112 to the network target 108. The intelligent detection system 102 may execute a comparison of the test response to the control response.

Abstract: In one embodiment, a human interactive proof portal 140 may use a biometric input to determine whether a user is a standard user or a malicious actor. The human interactive proof portal 140 may receive an access request 302 for an online data service 122 from a user device 110. The human interactive proof portal 140 may send a proof challenge 304 to the user device 110 for presentation to a user. The human interactive proof portal 140 may receive from the user device 110 a proof response 306 having a biometric metadata description 430 based on a biometric input from the user.

Abstract: In one embodiment, an intelligent detection system 102 may determine if a network target 108 is an adversarial site based on comparing responses to different network sources. The intelligent detection system 102 may select a test apparent network source 110 and a control apparent network source 112 from a network source pool 106. The intelligent detection system 102 may receive the test response responding to a test request from the test apparent network source 110 to a network target 108. The intelligent detection system 102 may receive the control response responding to a control request from the control apparent network source 112 to the network target 108. The intelligent detection system 102 may execute a comparison of the test response to the control response.

Abstract: The present invention concerns regulation of complement activation, in particular the fluid phase regulation of complement activation, and provides molecules comprising at least complement control protein modules 1-4 of complement factor H, DNA molecules encoding same, their use in the manufacture of a medicament for inhibiting complement activation and methods of same, together with DNA sequences encoding rat FH 4.3 and 1.0 kb mRNA.