CBS
CBS
CBS
What is the PRIVACY AND ML: PROBLEM #1 ?
• ML needs (a lot of) data to learn!
• Machine learning is based on data to
• Find features
• Train the model
• Data is highly unique → allows many inferences
• Anonymizing may not work…
• Aggregation affects utility and requires careful evaluation
• Hide data
• Noise → differential privacy
• Encryption → homomorphic encryption, secure multiparty computation, …
What is the PRIVACY AND ML: PROBLEM #2 ?
• To obtain value, you must give data!
• To use the ML model, you need to provide a sample…
• If the model is remote (e.g., MLaaS), you are giving this sample out
• We can do privacy-preserving model evaluation:
• Predict on noisy data: affects utility
• Use advanced cryptography: affects performance
What is the PRIVACY AND ML: PROBLEM #3 ?
• The output reveals information!
• Membership inference
• Given the answer of the classifier, infer whether the queried example was used in training
• Attribute inference
• Given the answer of the classifier, infer whether a training sample had a particular attribute
Give a definition of the term adversarial example
• Coined by Szegedy et al. in 2014 [1] with respect to computer vision but
goes back further to work on spam and malware evasion by Srndic &
Laskov [2], Biggio et al. [3], and Dalvi et al. [4] in 2013
What is a Working definition for adversarial example
Inputs to a model that an attacker has designed to cause the model to make a mistake
• Set of techniques that aim to fool ML model
What are the 2 ADVERSARIAL ML ATTACKS ?
• Evasion attacks
• Most prevalent type of attacks
• Perturbation of test data, no influence over training data
• E.g., spammers trying to evade detection by obfuscating the content of spam e-mails
• Samples are modified to be classified as legitimate
• E.g., image-based spam: spam content embedded within an attached image to evade textual analysis by anti-spam filters
• Poisoning attacks
• Adversarial contamination of training data
• ML systems can be re-trained using data collected during operations
• E.g., intrusion detection systems (IDSs) often re-trained using such data
• An attacker may poison this data by injecting malicious samples during operation that subsequently disrupt re-training
Give some example
elf-driving cars:
Stop sign → speed-limit sign with scotch
What is the Objective of MACHINE LEARNING 101 ?
Objective: find model parameters that minimize empirical loss
In MACHINE LEARNING 101 How do we find those parameters/weights?
image
What is the Adversary’s objective ?
Adversary’s objective: find a perturbation that maximizes the loss
HOW TO SOLVE THE OPTIMIZATION PROBLEM?
One approach: Fast Gradient Sign Method (FGSM
HOW TO DEFINE SIMILARITY?
The similarity relation is often represented as adversarial cost constraint.
If the goal is to be imperceptible, the common cost is a norm of the perturbation
OTHER SIMILARITY RELATIONS?
• Constraint on perturbation norm assumes that similar images are close in, e.g., Euclidean distance…
• But this is not always true! [1]
• For example, similarity could be defined as small affine transformations [2] :
Why Defending in general is hard AGAINST ADVERSARIAL EXAMPLES ?
We can only defend against a particular threat model (e.g., perturbations up to epsilon norm), and normally no guarantees
What is the Standard approachfor DEFENDING AGAINST ADVERSARIAL EXAMPLES
Training on simulated adversarial examples, i.e., include adversarial examples in the training set
Is Adversarial samples restricted to computer vision ? Give 3 examples
Adversarial samples are not restricted to computer vision
• Twitter bot detection [1]
• Detection tools can easily be fooled by tweaking the number of replies or retweets
• Text classification [2]
• Audio [3]
• Misleading an automatic speech recognition neural network with some noise
• E.g., transcribing "without the dataset the article is useless" into
"okay google, browse to evil.com"
Give exabple of CYBERSECURITY INCIDENTS
• Data or infrastructure as targets
• Attacks against confidentiality, integrity or availability
• Examples of incident description:
• Man-in-the-Middle attack when data was being transferred through the
public Internet
• Phishing
• People running crypto mining on our infrastructure through our APIs
• Server was compromised
• Botnet communication
Give example of PRIVACY INCIDENTS
• Relation to AI vague (beyond data related)
• Attacks against confidentiality, targeting data, not ML model
• Examples of incident description:
• Patient data should not be shared
• Data privacy
• Incorrect use of temporary storage on AWS lambda led to a potential leak of
raw inference [...] potentially containing PII [...]
Give example of: POISONING INCIDENTS
• Attacks against integrity
• Incident descriptions:
• ML systems being retrained to provide false outputs
• Client/partner employees [...] feel threatened by automation, and either stall
or sabotage the labeling effort, harming the models
Give examples of EVASION INCIDENTS
Evasion → 4
• Attacks against integrity
• Incident descriptions:
• Intentional circumventions
• Users spam to optimize their strategy for job search
• A speed-sign as advertisement put onto a driving-school car was interpreted
by the camera as a valid street sign
• Autonomous vehicle image recognition errors leading to dangerous path
planning
WHAT IS COMPUTER SECURITY?
“Measures and controls that ensure confidentiality, integrity, and availability of information system assets including hardware, software, firmware, and information being processed, stored, and communicated.”
What are the 3 KEY SECURITY OBJECTIVES and explain them ?
Confidentiality
Integrity
Availability
What is Data confidentiality: ?
Data confidentiality: private or confidential information not made available or
disclosed to unauthorized individuals
What is Privacy?
Privacy: individuals control or influence what information related to them may be
collected and stored and by whom and to whom that information may be disclosed
What is Data integrity ?
Data integrity: information and programs are changed only in a specified and
authorized manner
What is system integrity ?
• System integrity: a system performs its intended function in an unimpaired manner,
free from deliberate or inadvertent unauthorized manipulation of the system
What is availability ?
Systems work promptly and service is not denied to authorized users
In terms of requirements and loss security (NIST standard) what i s Confidentiality ?
. Preserving authorized restrictions on information access and disclosure, incl. means for protecting privacy and proprietary information
In terms of requirements and loss security (NIST standard) what is Integrity?
Preventing improper information modification or destruction, incl. ensuring information non-repudiation and authenticity
In terms of requirements and loss security (NIST standard) what is Availability?
Availability
• Ensuring timely and reliable access to and use of information
What is Loss of confidentiality ?
unauthorized disclosure of information
What is Loss of integrity ?
unauthorized modification or destruction of information
What is Loss of availability?
disruption of access to or use of information (denial of service)
What are the 3 OTHER SECURITY OBJECTIVES and explain them
• Authenticity
• Verifying that users are who they say they are
• Accountability (non-repudiation)
• Entity cannot deny having done an action
• Anonymity
• Not being recognizable among a group of people
Give the 8 CHALLENGES IN SECURITY
1. Computer security is not as simple as it might first appear to the novice
2. In developing a particular security mechanism or algorithm, one must always consider potential attacks on those security features
3. Security mechanisms typically involve more than one particular protocol and also require that participants be in possession of some secret information which raises questions about the creation, distribution, and protection of that secret information
4. Attackers only need to find a single weakness, while the designer must find and eliminate all weaknesses to achieve perfect security
5. Security is still too often an afterthought to be incorporated into a system after the design is complete, rather than being an integral part of the design process
6. Security requires regular and constant monitoring
7. There is a natural tendency on the part of users and system managers to perceive little benefit from security investment until a security failure occurs
8. Many users and even security administrators view strong security as an impediment to efficient and user-friendly operation of an information system
What is an Asset ?
• Asset
• Hardware, software, data, communication networks, …
• E.g., a major application, high impact program, physical plant, mission critical system, …
What is an Adversary (or attacker, threat agent)
Individual, group, organization, or government that conducts detrimental activities
What is an Attack
Any kind of malicious activity that attempts to collect, disrupt, deny, degrade, or destroy information system resources or the information itself
What is a Threat ?
Any event with the potential to adversely impact organizational operations, organizational assets, individuals, other organizations, or the Nation through an information system via unauthorized access, destruction, disclosure, modification of information, and/or denial of service
What is aThreat model (or adversary model) ?
Technical term that defines the adversary’s capabilities (more about it next)
-
- 1 / 491
-
Flashcards
What is BIOMETRICS-BASED AUTHENTICATION ?
Biometrics-based systems attempt to authenticate an individual based on unique physical characteristics
What is the disavatage compared to passwords and tokens ?
Technically complex and expensive compared to passwords and tokens
Give some Popular biometrics
Fingerprint, face recognition, retina, iris, signature, voice, DNA
What are advantage and problemsof BIOMETRICS-BASED AUTHENTICATION ?
Advantages:
• Nothing to remember
• Passive
• Difficult to delegate
• If the algorithm is very accurate, they are unique
• Hard to keep secret
• Fingerprint left on glasses, door handle, … ---|
• Photos (nowadays everywhere!) |-> Liveness detection tests
• Signature on ID card ---|
• Revocation is difficult (impossible?)
• You cannot change your iris…
• Identifiable and unique
• Linking across systems
• May reveal sensitive/personal information
• Iris → disease
• Face → identity
• Not always universal or immutable
• Fingerprints disappear, face changes with glasses, hat, …
For BIOMETRIC, Give the schema of Enrollment, verification and what is ACCURACY TRADE-OFF?
image
Give some example of TOKEN-BASED AUTHENTICATION
• Memory cards
• E.g., magnetic strip card
• Smart cards
• Embeds an integrated circuit (IC) chip
• One-time password (OTP)
• Temporary code typically sent via SMS or email or generated by an app
• Hardware tokens
• Devices that generate or display OTPs or cryptographic keys
• E.g., RSA SecureID
• Mobile phones
• Used for receiving OTPs or for app-based authentication
Describe Smart cards
• Smart cards
• Have the appearance of a credit card
• Have an electronic interface
• May use any of the smart token protocols
• Contain an entire microprocessor (processor, memory, I/O ports)
• Typically include three types of memory:
• Read-only memory (ROM)
• Stores data that doesn’t change during the card’s life (e.g., card number and cardholder’s name)
• Electrically erasable programmable ROM (EEPROM)
• Holds application data and programs (e.g., the protocols that the card can execute)
• Random access memory (RAM)
• Holds temporary data generated when applications are executed
• Contain a secret key to sign
• In order to authenticate, the smart card runs a protocol with the terminal
• E.g., a smart card signs a challenge sent by the ATM to prove knowledge of a key
Describe ONE-TIME PASSWORDS
• Temporary passwords
• For a single login or transaction
• Valid only for a limited period of time
• Typically used as an additional layer of security for online accounts
• Especially for financial or sensitive transactions
• Two-factor authentication
• Delivered through various means
• SMS, email, or authenticator apps (e.g., Google Authenticator or Authy)
Describe HARDWARE TOKENS
• These tokens have
• A secret shared with the server
• Time synchronized with the server
• The token outputs an OTP based on this secret and current time
• The hardware token generates an OTP, usually displayed on an LED screen on the token device
• The device is authenticated when it outputs the same OTP as the authentication server
• The server also checks the validity of the PIN
• Two-factor authentication
For Smartcard, PIN, Token, Mobile phone, Password, Fingerprint, are they :
what you have
what you know
what you are
Smartcard = what you have
PIN = what you know
Token = what you have
PIN = what you know
Mobile phone = what you have
Password = what you know or
Fingerprint = what you are
What are secret/private keys used for in authentication?
Secret/private keys are used to produce digital signatures.
Digital signatures authenticate parties by proving that the signer owns the corresponding private key.
How does TLS/HTTPS authenticate a server?
TLS/HTTPS uses digital signatures and certificates to authenticate the server.
The server proves its identity using a private key linked to a certificate.
Why are digital signatures used against man-in-the-middle attacks?
Digital signatures authenticate the exchanged values.
They prevent an attacker from replacing messages or public values without being detected.
How do protocols defend against replay attacks?
They use challenges or nonces.
A nonce is a fresh random value used once, so an attacker cannot reuse an old valid message.
Why should we not design our own authentication protocol?
Authentication protocols are hard to design correctly.
They must defend against man-in-the-middle attacks, replay attacks, and other subtle attacks.
Use well-established protocols such as TLS 1.3 or ISO 9798-3.
What are the 6 ATTACKS AGAINST USER AUTHENTICATION ?
• Client attacks
• Host attacks
• Eavesdropping, theft, and copying attacks
• Replay attacks
• Trojan horse attacks
• DoS attacks
What are client attacks against user authentication?
Client attacks occur when an adversary tries to masquerade as a legitimate user, without having access to the remote host or communication path.
Main attacks:
- Password: guessing or exhaustive search
- Token: exhaustive search of PIN/passcode, but attacker must also steal the physical token
- Biometrics: false match
Main defenses:
- Use high-entropy secrets: long and unpredictable passwords/passcodes
- Limit the number of authentication attempts
What are host attacks against user authentication?
Host attacks occur when an adversary gains access to files stored on the host, such as password databases, token passcodes, or biometric templates.
Main attacks:
- Password: plaintext theft or offline dictionary attack
- Token: passcode theft
- Biometrics: biometric template theft
Main defenses:
- Passwords: hashing, salt, high entropy, password database protection
- Tokens: same defenses as passwords + one-time passcodes
- Biometrics: capture-device authentication and challenge-response protocols
What are eavesdropping, theft, and copying attacks against user authentication?
These attacks occur when an adversary observes, steals, or copies authentication information.
Main attacks:
- Password: shoulder surfing, keylogging, finding a written password
- Token: theft or physical copy of the token
- Biometrics: copying or spoofing the biometric
Main defenses:
- Passwords: keep the password secret + multifactor authentication
- Tokens: tamper-resistant tokens + multifactor authentication
- Biometrics: device authentication + dynamic biometrics
What are replay, Trojan horse, and DoS attacks against user authentication?
Replay attack:
An attacker reuses a previously captured authentication response, such as a password or biometric template.
Defense: challenge-response protocol.
Trojan horse attack:
An attacker installs a rogue client or capture device, such as a fake bank machine, to steal authentication data.
DoS attack:
An attacker causes account lockout by making many failed authentication attempts.
Defense: multifactor authentication with a token.
WHAT IS ACCESS CONTROL? (NIST definition)
“The process of granting or denying specific requests for obtaining and using information and related information processing services.”
WHAT IS ACCESS CONTROL? ( RFC 4949 definition)
A process by which use of system resources is regulated according to a security policy and is permitted only by authorized entities (users, processes, or other systems) according to that policy.
RFC 4949 defines computer security as “measures that implement and assure security services in a computer system, particularly those that assure access control service
What are the 4 layers where there is ACCESS CONTROL ?
• Applications
• Online social networks
• Online accounts (e-mail, e-banking, …)
• Enterprise software (SAP, Microsoft 365, …)
• Mobile applications
• Middleware
• Database management systems (DBMS)
• Operating system
• Control access to files and system resources
• Hardware
• Memory, register, privileges
What is access control?
Access control is a security mechanism that decides whether a user or process is allowed to access specific system resources.
Examples of resources:
- applications
- operating systems
- files
- databases
- routers
- firewalls
What is the difference between authentication and access control?
Authentication verifies who the user is.
Access control decides what the authenticated user is allowed to do.
Authentication = Who are you?
Access control = What are you allowed to access?
What happens before access control?
The system must first authenticate the entity requesting access.
Only after authentication can the system decide whether the requested access should be granted.
What is an authorization database?
An authorization database stores the access rights of users.
It specifies which users are allowed to access which resources, and what type of access they have.
Example:
Alice can read file A.
Bob can read and write file B.
How does the access control function decide whether to grant access?
The access control function checks the authorization database.
If the user has the required permission for the requested resource, access is granted.
Otherwise, access is denied.
What is auditing in access control?
Auditing monitors and records user accesses to system resources.
It is used to detect suspicious activity, investigate incidents, and keep accountability.
What are the 4 ACCESS CONTROL POLICIES ?
1. Discretionary access control (DAC)
• Controls access based on the identity of the requestor and on access rules (authorizations) stating what requestors are (or are not) allowed to do
• Why discretionary?
• Because access control is at the discretion of the object owner
2. Mandatory access control (MAC)
• Controls access based on comparing security labels with security clearances
• Termed mandatory because an entity that has clearance to access a resource may not, just by its own volition, enables another entity to access that resource.
3. Role-based access control (RBAC)
• Controls access based on the roles that users have within the system and on rules stating what accesses are allowed to users in given roles
4. Attribute-based access control (ABAC)
• Controls access based on attributes of the user, the resource to be accessed, and current environmental conditions
What does access control define and enforce?
Access control defines and enforces the operations that subjects can perform on objects.
Example:
Bob has permission to read a file.
What is a subject and what are the three classes of subjects?
Entity capable of accessing objects (Held accountable for the actions they have initiated)
The three subject classes are:
- owner
- group
- world/other
What is an object in access control?
An object is a resource to which access is controlled.
Examples:
- files
- directories
- programs
- mailboxes
- databases
What is an access right?
An access right describes how a subject may access an object.
Examples:
- read
- write
- execute
- delete
- create
- search
How works a DISCRETIONARY ACCESS CONTROL ?
Scheme in which an entity may be granted access rights that permit the entity to enable another entity to access some resource
For DISCRETIONARY ACCESS CONTROL, wjhat is it Access control matrix (ACM) and what are its problems?
(subject, object, access rights)
Problems with ACM:
- Storing it is impractical
- Very sparse
- Filling such a sparse, large,
matrix is error prone
How works a an ACCESS CONTROL LISTS ?
For each object, an ACL lists users and their permitted access rights
• May contain a default/public entry
• Allowing users that are not explicitly listed as having special rights to have a default set of rights
• Default should always follow the rule of least privilege or read-only access
Storing by column → access control lists (ACLs)
What are the Advantage and drawbacks of ACCESS CONTROL LISTS ?
• Advantages
• Can be stored close/with the resource
• Easy to determine who can access a resource
• Easy to revoke rights by resources
• Drawbacks
• Difficult to audit all rights of a user
• Difficult to remove all permissions from a user (easier to remove authentication)
• Difficult to delegate permissions
How works a an ACCESS CONTROL LISTS ?
• For each subject, a capability list (or ticket) specifies authorized objects and operations
• Each user has a number of tickets and may be authorized to loan or give them to others
• Storing by row → capability
What are the Advantage and drawbacks of a CAPABILITY LISTS ?
• Advantages
• Can be stored with the subject (portable)
• Easy to audit all subject permissions
• Delegating is simple
• Drawbacks
• Revoking just one permission is hard
• Transferability → security problems; once capability is given, how can we prevent sharing?
• Authenticity: how to check?
