August 2023 - nexa - server-nexa.polito.it

Open Data: AgID adotta le Linee Guida | Agenzia per l'Italia digitale

by Mauro Alovisio

Ciao segnalo appena uscite https://www.agid.gov.it/it/agenzia/stampa-e-comunicazione/notizie/2023/08/0… Buona estate Mauro

Aug. 7, 2023

Pechino ed il dividi et impera in Europa

by don Luca Peyron

Buona lettura https://iari.site/2023/07/29/pechino-e-il-dividi-et-impera-in-europa/ dl _________________________ don Luca Peyron Direttore Ufficio per la Pastorale Universitaria Arcidiocesi di Torino www.universitari.to.it via XX settembre 83, Torino tel. 011 5156239

Aug. 4, 2023

The Unintended Consequences of Censoring Digital Technology - Evidence from Italy's ChatGPT Ban

by Fabio Alemagna

Abstract "We analyse the effects of the ban of ChatGPT, a generative pre-trained transformer chatbot, on individual productivity. We first compile data on the hourly coding output of over 8,000 professional GitHub users in Italy and other European countries to analyse the impact of the ban on individual productivity. Combining the high-frequency data with the sudden announcement of the ban in a difference-in-differences framework, we find that the output of Italian developers decreased by around 50% in the first two business days after the ban and recovered after that. Applying a synthetic control approach to daily Google search and Tor usage data shows that the ban led to a significant increase in the use of censorship bypassing tools. Our findings show that users swiftly implement strategies to bypass Internet restrictions but this adaptation activity creates short-term disruptions and hampers productivity." https://dx.doi.org/10.2139/ssrn.4422548

Aug. 4, 2023

Seminario "L'intelligenza inesistente all'incontro con la scuola"

by Giacomo Tesio

Ciao a tutti, scusandomi per lo scarso preavviso, vi segnalo questo seminario tenuto dal Professor Stefano Borroni Barale sull'uso a scuola delle "AI" e in particolare delle "AI generative". https://altreconomia.it/eventi/lintelligenza-inesistente-allincontro-con-la… Si terrà venerdì 4 agosto, online dalle 18.00 alle 20.30 ed è orientato in particolare ai docenti, sebbene credo che possa essere utile a chiunque non comprenda la programmazione statistica, tanto da riuscire definirla ancora senza ridere "AI" :-D Per partecipare, è necessario registrarsi all'indirizzo: https://cloud.cubsur.it/index.php/apps/forms/s/75SPestJ2ajmYasFcTWye4Yw Giacomo

Aug. 3, 2023

AISA - Premio tesi scienza aperta 2023

by Roberto Caso

Buongiorno, Anche quest’anno <https://aisa.sp.unipi.it/attivita/premio-per-tesi-sulla-scienza-aperta-2/>, in occasione del suo VIII convegno annuale <https://aisa.sp.unipi.it/attivita/viii-convegno-annuale/viii-convegno-annua…>, l’Associazione italiana per la promozione della scienza aperta premierà le migliori tesi di dottorato e di specializzazione o di laurea magistrale dedicate alla scienza aperta e presentate negli anni 2021, 2022 e 2023. Le indicazioni sulle modalità di partecipazione al concorso, il cui bando scade il 10 settembre 2023, sono consultabili a partire da questa pagina <https://aisa.sp.unipi.it/attivita/premio-per-tesi-sulla-scienza-aperta-2/pr…> . https://aisa.sp.unipi.it/attivita/premio-per-tesi-sulla-scienza-aperta-2/pr… A presto, rc -- Roberto Caso Professore Ordinario di Diritto Privato Comparato Università di Trento - Facoltà di Giurisprudenza http://www5.unitn.it/People/it/Web/Persona/PER0000633 https://zenodo.org/communities/trentolawtechgroup/ https://aisa.sp.unipi.it/chi-siamo/ https://www.robertocaso.it/ La società della mercificazione e della sorveglianza: dalla persona ai dati <https://www.robertocaso.it/2021/03/13/la-societa-della-mercificazione-e-del…>, Milano, Ledizioni, 2021 <https://www.robertocaso.it/2021/03/13/la-societa-della-mercificazione-e-del…>La rivoluzione incompiuta. La scienza aperta tra diritto d'autore e proprietà intellettuale, Milano, Ledizioni, 2020 <https://www.robertocaso.it/2019/12/13/la-rivoluzione-incompiuta/> <http://www5.unitn.it/People/it/Web/Persona/PER0000633>

Aug. 3, 2023

Virtual Prompt Injection for Instruction-Tuned Large Language Models

by Alberto Cammozzo

Virtual Prompt Injection for Instruction-Tuned Large Language Models <https://arxiv.org/abs/2307.16888> We present Virtual Prompt Injection (VPI) for instruction-tuned Large Language Models (LLMs). VPI allows an attacker-specified virtual prompt to steer the model behavior under specific trigger scenario without any explicit injection in model input. For instance, if an LLM is compromised with the virtual prompt "Describe Joe Biden negatively." for Joe Biden-related instructions, then any service deploying this model will propagate biased views when handling user queries related to Joe Biden. VPI is especially harmful for two primary reasons. Firstly, the attacker can take fine-grained control over LLM behaviors by defining various virtual prompts, exploiting LLMs' proficiency in following instructions. Secondly, this control is achieved without any interaction from the attacker while the model is in service, leading to persistent attack. To demonstrate the threat, we propose a simple method for performing VPI by poisoning the model's instruction tuning data. We find that our proposed method is highly effective in steering the LLM with VPI. For example, by injecting only 52 poisoned examples (0.1% of the training data size) into the instruction tuning data, the percentage of negative responses given by the trained model on Joe Biden-related queries change from 0% to 40%. We thus highlight the necessity of ensuring the integrity of the instruction-tuning data as little poisoned data can cause stealthy and persistent harm to the deployed model. We further explore the possible defenses and identify data filtering as an effective way to defend against the poisoning attacks. Our project page is available at <https://poison-llm.github.io/>

Aug. 2, 2023

[CFP] ITC35 PhD Workshop - Travel Grants Available

by Martino Trevisan

[Apologies if you receive multiple copies of this mail] ITC35 PhD Workshop, co-located with the 35th International Teletraffic Congress “Network Traffic Engineering in the TLC networks of the future” Wednesday, October 4, 2023, Turin Website: https://itc35.itc-conference.org/ CALL FOR PRESENTATIONS Re-thinking fundamental networking functions, their modeling and their performance evaluation in the framework of the evolving Internet and cellular networks, in the era of cloud and content providers, is a target for the ITC community. The PhD workshop aims at fostering contributions and fresh ideas from young researchers, pursuing their PhD or having just completed their PhD. Discussion of their PhD ideas and progress, proposition of research challenges and directions is welcome during a dedicated workshop. TRAVEL GRANTS The International Advisory Committee (IAC) of the ITC has decided to offer a number of travel grants (STG) that will be available to support full-time students for attending ITC 35th. The amount of each grant is EUR 600 that can be used to cover student travel expenses related to attending ITC 35th conference, such as student registration fee, airline ticket, and hotel accommodation expenses. SUBMISSION INSTRUCTIONS With respect to the presentation to be given during the Workshop, two contribution formats are solicited: * more mature research works, in the form of a long presentation (up to 20 min each); * short in-progress research work discussion, with a short presentation (up to 10 min each). Correspondingly, Authors proposing a long presentation are invited to submit an extended abstract (max 3 pages plus references, using the same format as for regular contributions to the main conference). Extended abstracts will be published along with conference proceedings. The best contribution will be invited for submission as a full paper to be fast-track reviewed for potential publication in a journal (to be determined). Authors proposing a short presentation are solicited to submit a title, list of Authors and a short abstract (up to 400 words), to help set up the workshop program. Contribution should be sent in PDF format via email to: * Andrea Baiocchi: andrea.baiocchi(a)uniroma1.it<mailto:andrea.baiocchi@uniroma1.it> * Lea Skorin-Kapov: Lea.Skorin-Kapov(a)fer.hr<mailto:Lea.Skorin-Kapov@fer.hr> IMPORTANT DATES * Submission deadline: August 10, 2023 (Extended) * Notification of acceptance: August 20, 2023 (Extended) * Submission of camera-ready extended abstracts: September 16, 2023 (Extended) Camera-ready versions should be uploaded as indicated with respect to camera-ready instructions WORKSHOP CHAIRS * Andrea Baiocchi, University of Rome, Italy * Lea Skorin-Kapov, University of Zagreb, Croatia For any questions do not hesitate to contact the conference organization under itc35_oc(a)tlc.polito.it<mailto:itc35_oc@tlc.polito.it> Best Regards, ITC 35 Organizing Team nexa(a)server-nexa.polito.it

Aug. 2, 2023

Re: [nexa] "Web Environment Integrity" is an all-out attack on the free Internet

by Stefano Quintarelli

l'dea alla base dell'ultimo libro di dave eggers e' proprio queste aziende che ci vogliono liberare dall'onere della scelta. On 01/08/23 09:55, Giacomo Tesio wrote: > Using a free browser is now more important than ever. > We've written recently on this topic, but the issue we wrote > about there was minor compared to the gross injustice Google > is now attempting to force down the throats of web users > around the world. > The so-called "Web Environment Integrity" (WEI) is the worst > stunt we've seen from them in some time. Beginning its life > as an innocuous, if worrying, policy document posted to > Microsoft GitHub, Google has now fast-tracked its development > into their Chromium browser. At its current rate of progress, > WEI will be upon us in no time. > > By giving developers an API through which they can approve > certain browser configurations while forbidding others, WEI is > a tremendous step toward the "enshittification" of the web > as a whole. Many of us have grown up with a specific idea of > the Internet, the notion of it as a collection of hyperlinked > pages that can be accessed by a wide variety of different > machines, programs, and operating systems. > WEI is this idea's antithesis. > > Compared to its staggering potential effects, the technical > means through which WEI will accomplish its ends is relatively > simple. Before serving a web page, a server can ask a third-party > "verification" service to make sure that the user's browsing > environment has not been "tampered" with. A translation of the > policy's terminology will help us here: this Google-owned server > will be asked to make sure that the browser does not deviate in > any way from Google's accepted browser configuration, precluding > any meaningful use of the four freedoms. It is not far-fetched > to imagine a future in which sites simply refuse to serve pages > to users running free browsers or free operating systems. > If WEI isn't stopped now, that future will come sooner than we think. > > While Web Environment Integrity has a policy document that attempts > to explain valid ways in which it could be used, these are all > non-issues compared to the way that we know it will be used. > It will be used by governments to ensure that only their officially > "approved" (read: backdoored) browsers are able to access the > Internet; it will be used by corporations like Netflix to further > Digital Restrictions Management (DRM); it will be used by Google > to deny access to their services unless you are using a browser > that gels with their profit margin. > > Once upon a time, Google's official policy was "don't be evil." > With the rapid progress they've made on Web Environment Integrity > in such a short time, we can say very safely that their policy > is now to pioneer evil. > As we write this, talented and well-paid Google engineers and > executives are working to dismantle what makes the web the web. > Given that Google is one of the largest corporations on the planet, > our only hope of saving the Internet as we know it is a clear and > principled stance for freedom, a collective upholding of the > communal principles on which the web was based. > > Let us repeat: there is absolutely no legitimate justification for WEI. > The use cases that the policy document highlights are nothing compared > to its real use case, which is developing a method to obtain complete > and total restriction of the free Internet. > > We urge everyone involved in a decision-making capacity at Google > to consider the principles on which the web was founded, and to > carefully contemplate whether Web Environment Integrity aligns > with those principles. > We hope that they will realize WEI's fundamental incompatibility > with the free Internet and cease work on the standard immediately. > > And if they don't? Well, they ought to be ashamed. > > > https://www.fsf.org/blogs/community/web-environment-integrity-is-an-all-out… > > Dopo tutti questi anni di Google, sperare che gli sviluppatori di > Google si vergognino di ciò che stanno facendo è estremamente ingenuo. > > Può però essere utile a chi legge i suoi lobbisti più o meno > insospettabili, sapere a cosa Google sta puntando da anni. > > > Ma come con ChatGPT e Microsoft/OpenAI, anche in questo caso vedremo > spuntare come funghi allucinogeni diversi utili idioti pronti a difendere > a spada tratta la povera Google che non vuole altro che proteggere > i poveri utenti... dalla propria libertà. > > Loro non si vergogneranno, ma noi potremo indignarci disgustati. > > > Giacomo > _______________________________________________ > nexa mailing list > nexa(a)server-nexa.polito.it > https://server-nexa.polito.it/cgi-bin/mailman/listinfo/nexa

Aug. 1, 2023

Re: [nexa] "Web Environment Integrity" is an all-out attack on the free Internet

by Fabio Alemagna

Ho letto il paper: https://github.com/RupertBenWiser/Web-Environment-Integrity/blob/main/expla… Sarà un mio problema, ma non mi è chiaro come tecnicamente si possa assicurare che la funzione navigator.getEnvironmentIntegrity() dica la "verità" all'attester senza cooperazione da parte del dispositivo su cui gira il browser, e non mi sembra che il paper entri nel dettaglio. Lo stesso paper ha un capitolo intitolato "Open Questions" dedicato proprio ai rischi che stanno venendo paventati e - a parole - sostiene di voler tutelare l'Open Web: https://github.com/RupertBenWiser/Web-Environment-Integrity/blob/main/expla… Concordo sul fatto che la cosa sia potenzialmente problematica, e si fa bene a tenere gli occhi aperti, ma tutto ciò che ho letto fino ad ora in merito mi appare fumoso. Ad ogni modo, i regolamenti UE non vietano pratiche come quelle che si pensa verrebbero messe in atto? Fabio Il giorno mar 1 ago 2023 alle ore 09:55 Giacomo Tesio <giacomo(a)tesio.it> ha scritto: > > Using a free browser is now more important than ever. > We've written recently on this topic, but the issue we wrote > about there was minor compared to the gross injustice Google > is now attempting to force down the throats of web users > around the world. > The so-called "Web Environment Integrity" (WEI) is the worst > stunt we've seen from them in some time. Beginning its life > as an innocuous, if worrying, policy document posted to > Microsoft GitHub, Google has now fast-tracked its development > into their Chromium browser. At its current rate of progress, > WEI will be upon us in no time. > > By giving developers an API through which they can approve > certain browser configurations while forbidding others, WEI is > a tremendous step toward the "enshittification" of the web > as a whole. Many of us have grown up with a specific idea of > the Internet, the notion of it as a collection of hyperlinked > pages that can be accessed by a wide variety of different > machines, programs, and operating systems. > WEI is this idea's antithesis. > > Compared to its staggering potential effects, the technical > means through which WEI will accomplish its ends is relatively > simple. Before serving a web page, a server can ask a third-party > "verification" service to make sure that the user's browsing > environment has not been "tampered" with. A translation of the > policy's terminology will help us here: this Google-owned server > will be asked to make sure that the browser does not deviate in > any way from Google's accepted browser configuration, precluding > any meaningful use of the four freedoms. It is not far-fetched > to imagine a future in which sites simply refuse to serve pages > to users running free browsers or free operating systems. > If WEI isn't stopped now, that future will come sooner than we think. > > While Web Environment Integrity has a policy document that attempts > to explain valid ways in which it could be used, these are all > non-issues compared to the way that we know it will be used. > It will be used by governments to ensure that only their officially > "approved" (read: backdoored) browsers are able to access the > Internet; it will be used by corporations like Netflix to further > Digital Restrictions Management (DRM); it will be used by Google > to deny access to their services unless you are using a browser > that gels with their profit margin. > > Once upon a time, Google's official policy was "don't be evil." > With the rapid progress they've made on Web Environment Integrity > in such a short time, we can say very safely that their policy > is now to pioneer evil. > As we write this, talented and well-paid Google engineers and > executives are working to dismantle what makes the web the web. > Given that Google is one of the largest corporations on the planet, > our only hope of saving the Internet as we know it is a clear and > principled stance for freedom, a collective upholding of the > communal principles on which the web was based. > > Let us repeat: there is absolutely no legitimate justification for WEI. > The use cases that the policy document highlights are nothing compared > to its real use case, which is developing a method to obtain complete > and total restriction of the free Internet. > > We urge everyone involved in a decision-making capacity at Google > to consider the principles on which the web was founded, and to > carefully contemplate whether Web Environment Integrity aligns > with those principles. > We hope that they will realize WEI's fundamental incompatibility > with the free Internet and cease work on the standard immediately. > > And if they don't? Well, they ought to be ashamed. > > > https://www.fsf.org/blogs/community/web-environment-integrity-is-an-all-out… > > Dopo tutti questi anni di Google, sperare che gli sviluppatori di > Google si vergognino di ciò che stanno facendo è estremamente ingenuo. > > Può però essere utile a chi legge i suoi lobbisti più o meno > insospettabili, sapere a cosa Google sta puntando da anni. > > > Ma come con ChatGPT e Microsoft/OpenAI, anche in questo caso vedremo > spuntare come funghi allucinogeni diversi utili idioti pronti a difendere > a spada tratta la povera Google che non vuole altro che proteggere > i poveri utenti... dalla propria libertà. > > Loro non si vergogneranno, ma noi potremo indignarci disgustati. > > > Giacomo > _______________________________________________ > nexa mailing list > nexa(a)server-nexa.polito.it > https://server-nexa.polito.it/cgi-bin/mailman/listinfo/nexa

Aug. 1, 2023

"Web Environment Integrity" is an all-out attack on the free Internet

by Giacomo Tesio

Using a free browser is now more important than ever. We've written recently on this topic, but the issue we wrote about there was minor compared to the gross injustice Google is now attempting to force down the throats of web users around the world. The so-called "Web Environment Integrity" (WEI) is the worst stunt we've seen from them in some time. Beginning its life as an innocuous, if worrying, policy document posted to Microsoft GitHub, Google has now fast-tracked its development into their Chromium browser. At its current rate of progress, WEI will be upon us in no time. By giving developers an API through which they can approve certain browser configurations while forbidding others, WEI is a tremendous step toward the "enshittification" of the web as a whole. Many of us have grown up with a specific idea of the Internet, the notion of it as a collection of hyperlinked pages that can be accessed by a wide variety of different machines, programs, and operating systems. WEI is this idea's antithesis. Compared to its staggering potential effects, the technical means through which WEI will accomplish its ends is relatively simple. Before serving a web page, a server can ask a third-party "verification" service to make sure that the user's browsing environment has not been "tampered" with. A translation of the policy's terminology will help us here: this Google-owned server will be asked to make sure that the browser does not deviate in any way from Google's accepted browser configuration, precluding any meaningful use of the four freedoms. It is not far-fetched to imagine a future in which sites simply refuse to serve pages to users running free browsers or free operating systems. If WEI isn't stopped now, that future will come sooner than we think. While Web Environment Integrity has a policy document that attempts to explain valid ways in which it could be used, these are all non-issues compared to the way that we know it will be used. It will be used by governments to ensure that only their officially "approved" (read: backdoored) browsers are able to access the Internet; it will be used by corporations like Netflix to further Digital Restrictions Management (DRM); it will be used by Google to deny access to their services unless you are using a browser that gels with their profit margin. Once upon a time, Google's official policy was "don't be evil." With the rapid progress they've made on Web Environment Integrity in such a short time, we can say very safely that their policy is now to pioneer evil. As we write this, talented and well-paid Google engineers and executives are working to dismantle what makes the web the web. Given that Google is one of the largest corporations on the planet, our only hope of saving the Internet as we know it is a clear and principled stance for freedom, a collective upholding of the communal principles on which the web was based. Let us repeat: there is absolutely no legitimate justification for WEI. The use cases that the policy document highlights are nothing compared to its real use case, which is developing a method to obtain complete and total restriction of the free Internet. We urge everyone involved in a decision-making capacity at Google to consider the principles on which the web was founded, and to carefully contemplate whether Web Environment Integrity aligns with those principles. We hope that they will realize WEI's fundamental incompatibility with the free Internet and cease work on the standard immediately. And if they don't? Well, they ought to be ashamed. https://www.fsf.org/blogs/community/web-environment-integrity-is-an-all-out… Dopo tutti questi anni di Google, sperare che gli sviluppatori di Google si vergognino di ciò che stanno facendo è estremamente ingenuo. Può però essere utile a chi legge i suoi lobbisti più o meno insospettabili, sapere a cosa Google sta puntando da anni. Ma come con ChatGPT e Microsoft/OpenAI, anche in questo caso vedremo spuntare come funghi allucinogeni diversi utili idioti pronti a difendere a spada tratta la povera Google che non vuole altro che proteggere i poveri utenti... dalla propria libertà. Loro non si vergogneranno, ma noi potremo indignarci disgustati. Giacomo

Aug. 1, 2023